abs
stringlengths 45
62
| Download PDF
stringlengths 50
84
| OpenReview
stringlengths 42
42
| title
stringlengths 10
168
| url
stringlengths 45
62
| authors
stringlengths 9
704
| detail_url
stringlengths 45
62
| tags
stringclasses 1
value | abstract
stringlengths 415
5.03k
⌀ |
|---|---|---|---|---|---|---|---|---|
https://proceedings.mlr.press/v202/tang23d.html
|
https://proceedings.mlr.press/v202/tang23d/tang23d.pdf
|
https://openreview.net/forum?id=QFw5VRmoGw
|
Understanding Self-Predictive Learning for Reinforcement Learning
|
https://proceedings.mlr.press/v202/tang23d.html
|
Yunhao Tang, Zhaohan Daniel Guo, Pierre Harvey Richemond, Bernardo Avila Pires, Yash Chandak, Remi Munos, Mark Rowland, Mohammad Gheshlaghi Azar, Charline Le Lan, Clare Lyle, András György, Shantanu Thakoor, Will Dabney, Bilal Piot, Daniele Calandriello, Michal Valko
|
https://proceedings.mlr.press/v202/tang23d.html
|
ICML 2023
|
We study the learning dynamics of self-predictive learning for reinforcement learning, a family of algorithms that learn representations by minimizing the prediction error of their own future latent representations. Despite its recent empirical success, such algorithms have an apparent defect: trivial representations (such as constants) minimize the prediction error, yet it is obviously undesirable to converge to such solutions. Our central insight is that careful designs of the optimization dynamics are critical to learning meaningful representations. We identify that a faster paced optimization of the predictor and semi-gradient updates on the representation, are crucial to preventing the representation collapse. Then in an idealized setup, we show self-predictive learning dynamics carries out spectral decomposition on the state transition matrix, effectively capturing information of the transition dynamics. Building on the theoretical insights, we propose bidirectional self-predictive learning, a novel self-predictive algorithm that learns two representations simultaneously. We examine the robustness of our theoretical insights with a number of small-scale experiments and showcase the promise of the novel representation learning algorithm with large-scale experiments.
|
https://proceedings.mlr.press/v202/tang23e.html
|
https://proceedings.mlr.press/v202/tang23e/tang23e.pdf
|
https://openreview.net/forum?id=PJ1NSXrzuI
|
DoMo-AC: Doubly Multi-step Off-policy Actor-Critic Algorithm
|
https://proceedings.mlr.press/v202/tang23e.html
|
Yunhao Tang, Tadashi Kozuno, Mark Rowland, Anna Harutyunyan, Remi Munos, Bernardo Avila Pires, Michal Valko
|
https://proceedings.mlr.press/v202/tang23e.html
|
ICML 2023
|
Multi-step learning applies lookahead over multiple time steps and has proved valuable in policy evaluation settings. However, in the optimal control case, the impact of multi-step learning has been relatively limited despite a number of prior efforts. Fundamentally, this might be because multi-step policy improvements require operations that cannot be approximated by stochastic samples, hence hindering the widespread adoption of such methods in practice. To address such limitations, we introduce doubly multi-step off-policy VI (DoMo-VI), a novel oracle algorithm that combines multi-step policy improvements and policy evaluations. DoMo-VI enjoys guaranteed convergence speed-up to the optimal policy and is applicable in general off-policy learning settings. We then propose doubly multi-step off-policy actor-critic (DoMo-AC), a practical instantiation of the DoMo-VI algorithm. DoMo-AC introduces a bias-variance trade-off that ensures improved policy gradient estimates. When combined with the IMPALA architecture, DoMo-AC has showed improvements over the baseline algorithm on Atari-57 game benchmarks.
|
https://proceedings.mlr.press/v202/tang23f.html
|
https://proceedings.mlr.press/v202/tang23f/tang23f.pdf
|
https://openreview.net/forum?id=BhMyLk0YNy
|
Towards Understanding Generalization of Graph Neural Networks
|
https://proceedings.mlr.press/v202/tang23f.html
|
Huayi Tang, Yong Liu
|
https://proceedings.mlr.press/v202/tang23f.html
|
ICML 2023
|
Graph neural networks (GNNs) are widely used in machine learning for graph-structured data. Even though GNNs have achieved remarkable success in real-world applications, understanding their working mechanism in theory is still on primary stage. In this paper, we move towards this goal from the perspective of generalization. Specifically, with consideration of stochastic optimization, we establish high probability bounds of generalization gap and gradients for transductive learning algorithms. After that, we provide high probability bounds of generalization gap for popular GNNs and analyze the factors affecting their generalization capability. These theoretical results reveal how the network architecture impacts the generalization gap. Experiments on benchmark datasets validate the theoretical findings. Our results provide new insights into understanding generalization of GNNs.
|
https://proceedings.mlr.press/v202/tang23g.html
|
https://proceedings.mlr.press/v202/tang23g/tang23g.pdf
|
https://openreview.net/forum?id=UOgyaswHo7
|
Towards a better understanding of representation dynamics under TD-learning
|
https://proceedings.mlr.press/v202/tang23g.html
|
Yunhao Tang, Remi Munos
|
https://proceedings.mlr.press/v202/tang23g.html
|
ICML 2023
|
TD-learning is a foundation reinforcement learning (RL) algorithm for value prediction. Critical to the accuracy of value predictions is the quality of state representations. In this work, we consider the question: how does end-to-end TD-learning impact the representation over time? Complementary to prior work, we provide a set of analysis that sheds further light on the representation dynamics under TD-learning. We first show that when the environments are reversible, end-to-end TD-learning strictly decreases the value approximation error over time. Under further assumptions on the environments, we can connect the representation dynamics with spectral decomposition over the transition matrix. This latter finding establishes fitting multiple value functions from randomly generated rewards as a useful auxiliary task for representation learning, as we empirically validate on both tabular and Atari game suites.
|
https://proceedings.mlr.press/v202/tang23h.html
|
https://proceedings.mlr.press/v202/tang23h/tang23h.pdf
|
https://openreview.net/forum?id=64xmfTWt7X
|
VA-learning as a more efficient alternative to Q-learning
|
https://proceedings.mlr.press/v202/tang23h.html
|
Yunhao Tang, Remi Munos, Mark Rowland, Michal Valko
|
https://proceedings.mlr.press/v202/tang23h.html
|
ICML 2023
|
In reinforcement learning, the advantage function is critical for policy improvement, but is often extracted from a learned Q-function. A natural question is: Why not learn the advantage function directly? In this work, we introduce VA-learning, which directly learns advantage function and value function using bootstrapping, without explicit reference to Q-functions. VA-learning learns off-policy and enjoys similar theoretical guarantees as Q-learning. Thanks to the direct learning of advantage function and value function, VA-learning improves the sample efficiency over Q-learning both in tabular implementations and deep RL agents on Atari-57 games. We also identify a close connection between VA-learning and the dueling architecture, which partially explains why a simple architectural change to DQN agents tends to improve performance.
|
https://proceedings.mlr.press/v202/tang23i.html
|
https://proceedings.mlr.press/v202/tang23i/tang23i.pdf
|
https://openreview.net/forum?id=dxUciRMAaE
|
Defects of Convolutional Decoder Networks in Frequency Representation
|
https://proceedings.mlr.press/v202/tang23i.html
|
Ling Tang, Wen Shen, Zhanpeng Zhou, Yuefeng Chen, Quanshi Zhang
|
https://proceedings.mlr.press/v202/tang23i.html
|
ICML 2023
|
In this paper, we prove the representation defects of a cascaded convolutional decoder network, considering the capacity of representing different frequency components of an input sample. We conduct the discrete Fourier transform on each channel of the feature map in an intermediate layer of the decoder network. Then, we extend the 2D circular convolution theorem to represent the forward and backward propagations through convolutional layers in the frequency domain. Based on this, we prove three defects in representing feature spectrums. First, we prove that the convolution operation, the zero-padding operation, and a set of other settings all make a convolutional decoder network more likely to weaken high-frequency components. Second, we prove that the upsampling operation generates a feature spectrum, in which strong signals repetitively appear at certain frequencies. Third, we prove that if the frequency components in the input sample and frequency components in the target output for regression have a small shift, then the decoder usually cannot be effectively learned.
|
https://proceedings.mlr.press/v202/tang23j.html
|
https://proceedings.mlr.press/v202/tang23j/tang23j.pdf
|
https://openreview.net/forum?id=FdCeFHTXrS
|
Difference-in-Differences Meets Tree-based Methods: Heterogeneous Treatment Effects Estimation with Unmeasured Confounding
|
https://proceedings.mlr.press/v202/tang23j.html
|
Caizhi Tang, Huiyuan Wang, Xinyu Li, Qing Cui, Longfei Li, Jun Zhou
|
https://proceedings.mlr.press/v202/tang23j.html
|
ICML 2023
|
This study considers the estimation of conditional causal effects in the presence of unmeasured confounding for a balanced panel with treatment imposed at the last time point. To address this, we combine Difference-in-differences (DiD) and tree-based methods and propose a new identification assumption that allows for the violation of the (conditional) parallel trends assumption adopted by most existing DiD methods. Under this new assumption, we prove partial identifiability of the conditional average treatment effect on the treated group (CATT). Our proposed method estimates CATT through a tree-based causal approach, guided by a novel splitting rule that avoids model misspecification and unnecessary auxiliary parameter estimation. The splitting rule measures both the error of fitting observed data and the violation of conditional parallel trends simultaneously. We also develop an ensemble of multiple trees via gradient boosting to further enhance performance. Experimental results on both synthetic and real-world datasets validate the effectiveness of our proposed method.
|
https://proceedings.mlr.press/v202/taniguchi23a.html
|
https://proceedings.mlr.press/v202/taniguchi23a/taniguchi23a.pdf
|
https://openreview.net/forum?id=mb3aTAlshp
|
End-to-end Training of Deep Boltzmann Machines by Unbiased Contrastive Divergence with Local Mode Initialization
|
https://proceedings.mlr.press/v202/taniguchi23a.html
|
Shohei Taniguchi, Masahiro Suzuki, Yusuke Iwasawa, Yutaka Matsuo
|
https://proceedings.mlr.press/v202/taniguchi23a.html
|
ICML 2023
|
We address the problem of biased gradient estimation in deep Boltzmann machines (DBMs). The existing method to obtain an unbiased estimator uses a maximal coupling based on a Gibbs sampler, but when the state is high-dimensional, it takes a long time to converge. In this study, we propose to use a coupling based on the Metropolis-Hastings (MH) and to initialize the state around a local mode of the target distribution. Because of the propensity of MH to reject proposals, the coupling tends to converge in only one step with a high probability, leading to high efficiency. We find that our method allows DBMs to be trained in an end-to-end fashion without greedy pretraining. We also propose some practical techniques to further improve the performance of DBMs. We empirically demonstrate that our training algorithm enables DBMs to show comparable generative performance to other deep generative models, achieving the FID score of 10.33 for MNIST.
|
https://proceedings.mlr.press/v202/tanwisuth23a.html
|
https://proceedings.mlr.press/v202/tanwisuth23a/tanwisuth23a.pdf
|
https://openreview.net/forum?id=0ndiQEXIcW
|
POUF: Prompt-Oriented Unsupervised Fine-tuning for Large Pre-trained Models
|
https://proceedings.mlr.press/v202/tanwisuth23a.html
|
Korawat Tanwisuth, Shujian Zhang, Huangjie Zheng, Pengcheng He, Mingyuan Zhou
|
https://proceedings.mlr.press/v202/tanwisuth23a.html
|
ICML 2023
|
Through prompting, large-scale pre-trained models have become more expressive and powerful, gaining significant attention in recent years. Though these big models have zero-shot capabilities, in general, labeled data are still required to adapt them to downstream tasks. To overcome this critical limitation, we propose an unsupervised fine-tuning framework to directly fine-tune the model or prompt on the unlabeled target data. We demonstrate how to apply our method to both language-augmented vision and masked-language models, by aligning the discrete distributions extracted from the prompts and target data. To verify our approach’s applicability, we conduct extensive experiments on image classification, sentiment analysis, and natural language inference tasks. Across 13 image-related tasks and 15 language-related ones, the proposed approach achieves consistent improvements over the baselines. PyTorch code is available at https://github.com/korawat-tanwisuth/POUF.
|
https://proceedings.mlr.press/v202/tao23a.html
|
https://proceedings.mlr.press/v202/tao23a/tao23a.pdf
|
https://openreview.net/forum?id=VQiPMxKCdm
|
Dual Focal Loss for Calibration
|
https://proceedings.mlr.press/v202/tao23a.html
|
Linwei Tao, Minjing Dong, Chang Xu
|
https://proceedings.mlr.press/v202/tao23a.html
|
ICML 2023
|
The use of deep neural networks in real-world applications require well-calibrated networks with confidence scores that accurately reflect the actual probability. However, it has been found that these networks often provide over-confident predictions, which leads to poor calibration. Recent efforts have sought to address this issue by focal loss to reduce over-confidence, but this approach can also lead to under-confident predictions. While different variants of focal loss have been explored, it is difficult to find a balance between over-confidence and under-confidence. In our work, we propose a new loss function by focusing on dual logits. Our method not only considers the ground truth logit, but also take into account the highest logit ranked after the ground truth logit. By maximizing the gap between these two logits, our proposed dual focal loss can achieve a better balance between over-confidence and under-confidence. We provide theoretical evidence to support our approach and demonstrate its effectiveness through evaluations on multiple models and datasets, where it achieves state-of-the-art performance. Code is available at https://github.com/Linwei94/DualFocalLoss
|
https://proceedings.mlr.press/v202/tao23b.html
|
https://proceedings.mlr.press/v202/tao23b/tao23b.pdf
|
https://openreview.net/forum?id=SryTYOIGJx
|
Abstract-to-Executable Trajectory Translation for One-Shot Task Generalization
|
https://proceedings.mlr.press/v202/tao23b.html
|
Stone Tao, Xiaochen Li, Tongzhou Mu, Zhiao Huang, Yuzhe Qin, Hao Su
|
https://proceedings.mlr.press/v202/tao23b.html
|
ICML 2023
|
Training long-horizon robotic policies in complex physical environments is essential for many applications, such as robotic manipulation. However, learning a policy that can generalize to unseen tasks is challenging. In this work, we propose to achieve one-shot task generalization by decoupling plan generation and plan execution. Specifically, our method solves complex long-horizon tasks in three steps: build a paired abstract environment by simplifying geometry and physics, generate abstract trajectories, and solve the original task by an abstract-to-executable trajectory translator. In the abstract environment, complex dynamics such as physical manipulation are removed, making abstract trajectories easier to generate. However, this introduces a large domain gap between abstract trajectories and the actual executed trajectories as abstract trajectories lack low-level details and are not aligned frame-to-frame with the executed trajectory. In a manner reminiscent of language translation, our approach leverages a seq-to-seq model to overcome the large domain gap between the abstract and executable trajectories, enabling the low-level policy to follow the abstract trajectory. Experimental results on various unseen long-horizon tasks with different robot embodiments demonstrate the practicability of our methods to achieve one-shot task generalization.
|
https://proceedings.mlr.press/v202/taori23a.html
|
https://proceedings.mlr.press/v202/taori23a/taori23a.pdf
|
https://openreview.net/forum?id=8JXMDw2xGa
|
Data Feedback Loops: Model-driven Amplification of Dataset Biases
|
https://proceedings.mlr.press/v202/taori23a.html
|
Rohan Taori, Tatsunori Hashimoto
|
https://proceedings.mlr.press/v202/taori23a.html
|
ICML 2023
|
Datasets scraped from the internet have been critical to large-scale machine learning. Yet, its success puts the utility of future internet-derived datasets at potential risk, as model outputs begin to replace human annotations as a source of supervision. In this work, we formalize a system where interactions with one model are recorded as history and scraped as training data in the future. We then analyze its stability over time by tracking changes to a test-time bias statistic (e.g. gender bias of model predictions). We find that the degree of bias amplification is closely linked to whether the model’s outputs behave like samples from the training distribution, a behavior which we characterize and define as uniform faithfulness. Experiments in three conditional prediction scenarios – image classification, visual role-labeling, and language generation – demonstrate that models that exhibit a sampling-like behavior are more faithful and thus more stable. Based on this insight, we propose an intervention to help mitigate and stabilize unstable feedback systems.
|
https://proceedings.mlr.press/v202/tarun23a.html
|
https://proceedings.mlr.press/v202/tarun23a/tarun23a.pdf
|
https://openreview.net/forum?id=oJANAXYc18
|
Deep Regression Unlearning
|
https://proceedings.mlr.press/v202/tarun23a.html
|
Ayush Kumar Tarun, Vikram Singh Chundawat, Murari Mandal, Mohan Kankanhalli
|
https://proceedings.mlr.press/v202/tarun23a.html
|
ICML 2023
|
With the introduction of data protection and privacy regulations, it has become crucial to remove the lineage of data on demand from a machine learning (ML) model. In the last few years, there have been notable developments in machine unlearning to remove the information of certain training data efficiently and effectively from ML models. In this work, we explore unlearning for the regression problem, particularly in deep learning models. Unlearning in classification and simple linear regression has been considerably investigated. However, unlearning in deep regression models largely remains an untouched problem till now. In this work, we introduce deep regression unlearning methods that generalize well and are robust to privacy attacks. We propose the Blindspot unlearning method which uses a novel weight optimization process. A randomly initialized model, partially exposed to the retain samples and a copy of the original model are used together to selectively imprint knowledge about the data that we wish to keep and scrub off the information of the data we wish to forget. We also propose a Gaussian fine tuning method for regression unlearning. The existing unlearning metrics for classification are not directly applicable to regression unlearning. Therefore, we adapt these metrics for the regression setting. We conduct regression unlearning experiments for computer vision, natural language processing and forecasting applications. Our methods show excellent performance for all these datasets across all the metrics. Source code: https://github.com/ayu987/deep-regression-unlearning
|
https://proceedings.mlr.press/v202/teneggi23a.html
|
https://proceedings.mlr.press/v202/teneggi23a/teneggi23a.pdf
|
https://openreview.net/forum?id=UyBZ4zIOzV
|
How to Trust Your Diffusion Model: A Convex Optimization Approach to Conformal Risk Control
|
https://proceedings.mlr.press/v202/teneggi23a.html
|
Jacopo Teneggi, Matthew Tivnan, Web Stayman, Jeremias Sulam
|
https://proceedings.mlr.press/v202/teneggi23a.html
|
ICML 2023
|
Score-based generative modeling, informally referred to as diffusion models, continue to grow in popularity across several important domains and tasks. While they provide high-quality and diverse samples from empirical distributions, important questions remain on the reliability and trustworthiness of these sampling procedures for their responsible use in critical scenarios. Conformal prediction is a modern tool to construct finite-sample, distribution-free uncertainty guarantees for any black-box predictor. In this work, we focus on image-to-image regression tasks and we present a generalization of the Risk-Controlling Prediction Sets (RCPS) procedure, that we term $K$-RCPS, which allows to $(i)$ provide entrywise calibrated intervals for future samples of any diffusion model, and $(ii)$ control a certain notion of risk with respect to a ground truth image with minimal mean interval length. Differently from existing conformal risk control procedures, ours relies on a novel convex optimization approach that allows for multidimensional risk control while provably minimizing the mean interval length. We illustrate our approach on two real-world image denoising problems: on natural images of faces as well as on computed tomography (CT) scans of the abdomen, demonstrating state of the art performance.
|
https://proceedings.mlr.press/v202/tenenbaum23a.html
|
https://proceedings.mlr.press/v202/tenenbaum23a/tenenbaum23a.pdf
|
https://openreview.net/forum?id=pWeQdceMHL
|
Concurrent Shuffle Differential Privacy Under Continual Observation
|
https://proceedings.mlr.press/v202/tenenbaum23a.html
|
Jay Tenenbaum, Haim Kaplan, Yishay Mansour, Uri Stemmer
|
https://proceedings.mlr.press/v202/tenenbaum23a.html
|
ICML 2023
|
We introduce the concurrent shuffle model of differential privacy. In this model we have multiple concurrent shufflers permuting messages from different, possibly overlapping, batches of users. Similarly to the standard (single) shuffler model, the privacy requirement is that the concatenation of all shuffled messages should be differentially private. We study the private continual summation problem (a.k.a. the counter problem) and show that the concurrent shuffle model allows for significantly improved error compared to a standard (single) shuffler model. Specifically, we give a summation algorithm with error $\tilde{O}(n^{1/(2k+1)})$ with $k$ concurrent shufflers on a sequence of length $n$. Furthermore, we prove that this bound is tight for any $k$, even if the algorithm can choose the sizes of the batches adaptively. For $k=\log n$ shufflers, the resulting error is polylogarithmic, much better than $\tilde{\Theta}(n^{1/3})$ which we show is the smallest possible with a single shuffler. We use our online summation algorithm to get algorithms with improved regret bounds for the contextual linear bandit problem. In particular we get optimal $\tilde{O}(\sqrt{n})$ regret with $k= \tilde{\Omega}(\log n)$ concurrent shufflers.
|
https://proceedings.mlr.press/v202/teng23a.html
|
https://proceedings.mlr.press/v202/teng23a/teng23a.pdf
|
https://openreview.net/forum?id=PQgjker1cd
|
Finding Generalization Measures by Contrasting Signal and Noise
|
https://proceedings.mlr.press/v202/teng23a.html
|
Jiaye Teng, Bohang Zhang, Ruichen Li, Haowei He, Yequan Wang, Yan Tian, Yang Yuan
|
https://proceedings.mlr.press/v202/teng23a.html
|
ICML 2023
|
Generalization is one of the most fundamental challenges in deep learning, aiming to predict model performances on unseen data. Empirically, such predictions usually rely on a validation set, while recent works showed that an unlabeled validation set also works. Without validation sets, it is extremely difficult to obtain non-vacuous generalization bounds, which leads to a weaker task of finding generalization measures that monotonically relate to generalization error. In this paper, we propose a new generalization measure REF Complexity (RElative Fitting degree between signal and noise), motivated by the intuition that a given model-algorithm pair may generalize well if it fits signal (e.g., true labels) fast while fitting noise (e.g., random labels) slowly. Empirically, REF Complexity monotonically relates to test accuracy in real-world datasets without accessing additional validation sets, achieving -0.988 correlation on CIFAR-10 and -0.960 correlation on CIFAR-100. We further theoretically verify the utility of REF Complexity under three different cases, including convex and smooth regimes with stochastic gradient descent, smooth regimes (not necessarily convex) with stochastic gradient Langevin dynamics, and linear regimes with gradient descent. The code is available at https://github.com/962086838/REF-complexity.
|
https://proceedings.mlr.press/v202/tennenholtz23a.html
|
https://proceedings.mlr.press/v202/tennenholtz23a/tennenholtz23a.pdf
|
https://openreview.net/forum?id=rdOuTlTUMX
|
Reinforcement Learning with History Dependent Dynamic Contexts
|
https://proceedings.mlr.press/v202/tennenholtz23a.html
|
Guy Tennenholtz, Nadav Merlis, Lior Shani, Martin Mladenov, Craig Boutilier
|
https://proceedings.mlr.press/v202/tennenholtz23a.html
|
ICML 2023
|
We introduce Dynamic Contextual Markov Decision Processes (DCMDPs), a novel reinforcement learning framework for history-dependent environments that generalizes the contextual MDP framework to handle non-Markov environments, where contexts change over time. We consider special cases of the model, with a focus on logistic DCMDPs, which break the exponential dependence on history length by leveraging aggregation functions to determine context transitions. This special structure allows us to derive an upper-confidence-bound style algorithm for which we establish regret bounds. Motivated by our theoretical results, we introduce a practical model-based algorithm for logistic DCMDPs that plans in a latent space and uses optimism over history-dependent features. We demonstrate the efficacy of our approach on a recommendation task (using MovieLens data) where user behavior dynamics evolve in response to recommendations.
|
https://proceedings.mlr.press/v202/ter-minassian23a.html
|
https://proceedings.mlr.press/v202/ter-minassian23a/ter-minassian23a.pdf
|
https://openreview.net/forum?id=u8VEJNykA5
|
PWSHAP: A Path-Wise Explanation Model for Targeted Variables
|
https://proceedings.mlr.press/v202/ter-minassian23a.html
|
Lucile Ter-Minassian, Oscar Clivio, Karla Diazordaz, Robin J. Evans, Christopher C. Holmes
|
https://proceedings.mlr.press/v202/ter-minassian23a.html
|
ICML 2023
|
Predictive black-box models can exhibit high-accuracy but their opaque nature hinders their uptake in safety-critical deployment environments. Explanation methods (XAI) can provide confidence for decision-making through increased transparency. However, existing XAI methods are not tailored towards models in sensitive domains where one predictor is of special interest, such as a treatment effect in a clinical model, or ethnicity in policy models. We introduce Path-Wise Shapley effects (PWSHAP), a framework for assessing the targeted effect of a binary (e.g. treatment) variable from a complex outcome model. Our approach augments the predictive model with a user-defined directed acyclic graph (DAG). The method then uses the graph alongside on-manifold Shapley values to identify effects along causal pathways whilst maintaining robustness to adversarial attacks. We establish error bounds for the identified path-wise Shapley effects and for Shapley values. We show PWSHAP can perform local bias and mediation analyses with faithfulness to the model. Further, if the targeted variable is randomised we can quantify local effect modification. We demonstrate the resolution, interpretability and true locality of our approach on examples and a real-world experiment.
|
https://proceedings.mlr.press/v202/tewari23a.html
|
https://proceedings.mlr.press/v202/tewari23a/tewari23a.pdf
|
https://openreview.net/forum?id=uP5xXIULdH
|
On the Estimation of Gaussian Mixture Copula Models
|
https://proceedings.mlr.press/v202/tewari23a.html
|
Ashutosh Tewari
|
https://proceedings.mlr.press/v202/tewari23a.html
|
ICML 2023
|
This paper revisits Gaussian Mixture Copula Model (GMCM), a more expressive alternative to the widely used Gaussian Mixture Model (GMM), with the goal to make its parameter estimation tractable. Both the Expectation Maximization and the direct Likelihood Maximization frameworks for GMCM have to grapple with a likelihood function that lacks a closed form. This has led to a few approximation schemes that alleviate the problem, nonetheless leaving the issue still unresolved. Additionally, past works have alluded to an additional challenge of parameter non-identifiability, but none has offered a rigorous treatment and a commensurate solution framework to overcome the same. This work offers solutions to each of these issues in an attempt to help GMCM realize its full potential. The source of non-identifiability is not only proven but also suitable priors are proposed that eliminate the problem. Additionally, an efficient numerical framework is proposed to evaluate the intractable likelihood function, while also providing its analytical derivatives. Finally, a view of GMCM as a series of bijective mappings from a base distribution is presented, which paves the way to synthesize GMCM using modern, probabilistic programming languages (PPLs). The main claims of this work are supported by empirical evidence gathered on synthetic and real-world datasets.
|
https://proceedings.mlr.press/v202/thopalli23a.html
|
https://proceedings.mlr.press/v202/thopalli23a/thopalli23a.pdf
|
https://openreview.net/forum?id=K26zQKvXiR
|
Target-Aware Generative Augmentations for Single-Shot Adaptation
|
https://proceedings.mlr.press/v202/thopalli23a.html
|
Kowshik Thopalli, Rakshith Subramanyam, Pavan K. Turaga, Jayaraman J. Thiagarajan
|
https://proceedings.mlr.press/v202/thopalli23a.html
|
ICML 2023
|
In this paper, we address the problem of adapting models from a source domain to a target domain, a task that has become increasingly important due to the brittle generalization of deep neural networks. While several test-time adaptation techniques have emerged, they typically rely on synthetic toolbox data augmentations in cases of limited target data availability. We consider the challenging setting of single-shot adaptation and explore the design of augmentation strategies. We argue that augmentations utilized by existing methods are insufficient to handle large distribution shifts, and hence propose a new approach SiSTA, which first fine-tunes a generative model from the source domain using a single-shot target, and then employs novel sampling strategies for curating synthetic target data. Using experiments on a variety of benchmarks, distribution shifts and image corruptions, we find that SiSTA produces significantly improved generalization over existing baselines in face attribute detection and multi-class object recognition. Furthermore, SiSTA performs competitively to models obtained by training on larger target datasets. Our codes can be accessed at https://github.com/Rakshith-2905/SiSTA
|
https://proceedings.mlr.press/v202/tian23a.html
|
https://proceedings.mlr.press/v202/tian23a/tian23a.pdf
|
https://openreview.net/forum?id=tCCwr7SmAM
|
ELSA: Efficient Label Shift Adaptation through the Lens of Semiparametric Models
|
https://proceedings.mlr.press/v202/tian23a.html
|
Qinglong Tian, Xin Zhang, Jiwei Zhao
|
https://proceedings.mlr.press/v202/tian23a.html
|
ICML 2023
|
We study the domain adaptation problem with label shift in this work. Under the label shift context, the marginal distribution of the label varies across the training and testing datasets, while the conditional distribution of features given the label is the same. Traditional label shift adaptation methods either suffer from large estimation errors or require cumbersome post-prediction calibrations. To address these issues, we first propose a moment-matching framework for adapting the label shift based on the geometry of the influence function. Under such a framework, we propose a novel method named $\underline{\mathrm{E}}$fficient $\underline{\mathrm{L}}$abel $\underline{\mathrm{S}}$hift $\underline{\mathrm{A}}$daptation (ELSA), in which the adaptation weights can be estimated by solving linear systems. Theoretically, the ELSA estimator is $\sqrt{n}$-consistent ($n$ is the sample size of the source data) and asymptotically normal. Empirically, we show that ELSA can achieve state-of-the-art estimation performances without post-prediction calibrations, thus, gaining computational efficiency.
|
https://proceedings.mlr.press/v202/tiao23a.html
|
https://proceedings.mlr.press/v202/tiao23a/tiao23a.pdf
|
https://openreview.net/forum?id=s58a6Pxw7V
|
Spherical Inducing Features for Orthogonally-Decoupled Gaussian Processes
|
https://proceedings.mlr.press/v202/tiao23a.html
|
Louis C. Tiao, Vincent Dutordoir, Victor Picheny
|
https://proceedings.mlr.press/v202/tiao23a.html
|
ICML 2023
|
Despite their many desirable properties, Gaussian processes (GPs) are often compared unfavorably to deep neural networks (NNs) for lacking the ability to learn representations. Recent efforts to bridge the gap between GPs and deep NNs have yielded a new class of inter-domain variational GPs in which the inducing variables correspond to hidden units of a feedforward NN. In this work, we examine some practical issues associated with this approach and propose an extension that leverages the orthogonal decomposition of GPs to mitigate these limitations. In particular, we introduce spherical inter-domain features to construct more flexible data-dependent basis functions for both the principal and orthogonal components of the GP approximation and show that incorporating NN activation features under this framework not only alleviates these shortcomings but is more scalable than alternative strategies. Experiments on multiple benchmark datasets demonstrate the effectiveness of our approach.
|
https://proceedings.mlr.press/v202/tiapkin23a.html
|
https://proceedings.mlr.press/v202/tiapkin23a/tiapkin23a.pdf
|
https://openreview.net/forum?id=wcUppxYfLH
|
Fast Rates for Maximum Entropy Exploration
|
https://proceedings.mlr.press/v202/tiapkin23a.html
|
Daniil Tiapkin, Denis Belomestny, Daniele Calandriello, Eric Moulines, Remi Munos, Alexey Naumov, Pierre Perrault, Yunhao Tang, Michal Valko, Pierre Menard
|
https://proceedings.mlr.press/v202/tiapkin23a.html
|
ICML 2023
|
We address the challenge of exploration in reinforcement learning (RL) when the agent operates in an unknown environment with sparse or no rewards. In this work, we study the maximum entropy exploration problem of two different types. The first type is visitation entropy maximization previously considered by Hazan et al. (2019) in the discounted setting. For this type of exploration, we propose a game-theoretic algorithm that has $\widetilde{\mathcal{O}}(H^3S^2A/\varepsilon^2)$ sample complexity thus improving the $\varepsilon$-dependence upon existing results, where $S$ is a number of states, $A$ is a number of actions, $H$ is an episode length, and $\varepsilon$ is a desired accuracy. The second type of entropy we study is the trajectory entropy. This objective function is closely related to the entropy-regularized MDPs, and we propose a simple algorithm that has a sample complexity of order $\widetilde{\mathcal{O}}(\mathrm{poly}(S,A,H)/\varepsilon)$. Interestingly, it is the first theoretical result in RL literature that establishes the potential statistical advantage of regularized MDPs for exploration. Finally, we apply developed regularization techniques to reduce sample complexity of visitation entropy maximization to $\widetilde{\mathcal{O}}(H^2SA/\varepsilon^2)$, yielding a statistical separation between maximum entropy exploration and reward-free exploration.
|
https://proceedings.mlr.press/v202/tifrea23a.html
|
https://proceedings.mlr.press/v202/tifrea23a/tifrea23a.pdf
|
https://openreview.net/forum?id=evgruLDFtA
|
Margin-based sampling in high dimensions: When being active is less efficient than staying passive
|
https://proceedings.mlr.press/v202/tifrea23a.html
|
Alexandru Tifrea, Jacob Clarysse, Fanny Yang
|
https://proceedings.mlr.press/v202/tifrea23a.html
|
ICML 2023
|
It is widely believed that given the same labeling budget, active learning (AL) algorithms like margin-based active learning achieve better predictive performance than passive learning (PL), albeit at a higher computational cost. Recent empirical evidence suggests that this added cost might be in vain, as margin-based AL can sometimes perform even worse than PL. While existing works offer different explanations in the low-dimensional regime, this paper shows that the underlying mechanism is entirely different in high dimensions: we prove for logistic regression that PL outperforms margin-based AL even for noiseless data and when using the Bayes optimal decision boundary for sampling. Insights from our proof indicate that this high-dimensional phenomenon is exacerbated when the separation between the classes is small. We corroborate this intuition with experiments on 20 high-dimensional datasets spanning a diverse range of applications, from finance and histology to chemistry and computer vision.
|
https://proceedings.mlr.press/v202/tigas23a.html
|
https://proceedings.mlr.press/v202/tigas23a/tigas23a.pdf
|
https://openreview.net/forum?id=oR2IsISm1X
|
Differentiable Multi-Target Causal Bayesian Experimental Design
|
https://proceedings.mlr.press/v202/tigas23a.html
|
Panagiotis Tigas, Yashas Annadani, Desi R. Ivanova, Andrew Jesson, Yarin Gal, Adam Foster, Stefan Bauer
|
https://proceedings.mlr.press/v202/tigas23a.html
|
ICML 2023
|
We introduce a gradient-based approach for the problem of Bayesian optimal experimental design to learn causal models in a batch setting — a critical component for causal discovery from finite data where interventions can be costly or risky. Existing methods rely on greedy approximations to construct a batch of experiments while using black-box methods to optimize over a single target-state pair to intervene with. In this work, we completely dispose of the black-box optimization techniques and greedy heuristics and instead propose a conceptually simple end-to-end gradient-based optimization procedure to acquire a set of optimal intervention target-value pairs. Such a procedure enables parameterization of the design space to efficiently optimize over a batch of multi-target-state interventions, a setting which has hitherto not been explored due to its complexity. We demonstrate that our proposed method outperforms baselines and existing acquisition strategies in both single-target and multi-target settings across a number of synthetic datasets.
|
https://proceedings.mlr.press/v202/tiomoko23a.html
|
https://proceedings.mlr.press/v202/tiomoko23a/tiomoko23a.pdf
|
https://openreview.net/forum?id=prMTSnjVuR
|
PCA-based Multi-Task Learning: a Random Matrix Approach
|
https://proceedings.mlr.press/v202/tiomoko23a.html
|
Malik Tiomoko, Romain Couillet, Frederic Pascal
|
https://proceedings.mlr.press/v202/tiomoko23a.html
|
ICML 2023
|
The article proposes and theoretically analyses a computationally efficient multi-task learning (MTL) extension of popular principal component analysis (PCA)-based supervised learning schemes. The analysis reveals that (i) by default, learning may dramatically fail by suffering from negative transfer, but that (ii) simple counter-measures on data labels avert negative transfer and necessarily result in improved performances. Supporting experiments on synthetic and real data benchmarks show that the proposed method achieves comparable performance with state-of-the-art MTL methods but at a significantly reduced computational cost.
|
https://proceedings.mlr.press/v202/tirer23a.html
|
https://proceedings.mlr.press/v202/tirer23a/tirer23a.pdf
|
https://openreview.net/forum?id=HFATn6FFcG
|
Perturbation Analysis of Neural Collapse
|
https://proceedings.mlr.press/v202/tirer23a.html
|
Tom Tirer, Haoxiang Huang, Jonathan Niles-Weed
|
https://proceedings.mlr.press/v202/tirer23a.html
|
ICML 2023
|
Training deep neural networks for classification often includes minimizing the training loss beyond the zero training error point. In this phase of training, a "neural collapse" behavior has been observed: the variability of features (outputs of the penultimate layer) of within-class samples decreases and the mean features of different classes approach a certain tight frame structure. Recent works analyze this behavior via idealized unconstrained features models where all the minimizers exhibit exact collapse. However, with practical networks and datasets, the features typically do not reach exact collapse, e.g., because deep layers cannot arbitrarily modify intermediate features that are far from being collapsed. In this paper, we propose a richer model that can capture this phenomenon by forcing the features to stay in the vicinity of a predefined features matrix (e.g., intermediate features). We explore the model in the small vicinity case via perturbation analysis and establish results that cannot be obtained by the previously studied models. For example, we prove reduction in the within-class variability of the optimized features compared to the predefined input features (via analyzing gradient flow on the "central-path" with minimal assumptions), analyze the minimizers in the near-collapse regime, and provide insights on the effect of regularization hyperparameters on the closeness to collapse. We support our theory with experiments in practical deep learning settings.
|
https://proceedings.mlr.press/v202/tiwari23a.html
|
https://proceedings.mlr.press/v202/tiwari23a/tiwari23a.pdf
|
https://openreview.net/forum?id=DnTVBs6zbz
|
Overcoming Simplicity Bias in Deep Networks using a Feature Sieve
|
https://proceedings.mlr.press/v202/tiwari23a.html
|
Rishabh Tiwari, Pradeep Shenoy
|
https://proceedings.mlr.press/v202/tiwari23a.html
|
ICML 2023
|
Simplicity bias is the concerning tendency of deep networks to over-depend on simple, weakly predictive features, to the exclusion of stronger, more complex features. This causes biased, incorrect model predictions in many real-world applications, exacerbated by incomplete training data containing spurious feature-label correlations. We propose a direct, interventional method for addressing simplicity bias in DNNs, which we call the feature sieve. We aim to automatically identify and suppress easily-computable spurious features in lower layers of the network, thereby allowing the higher network levels to extract and utilize richer, more meaningful representations. We provide concrete evidence of this differential suppression & enhancement of relevant features on both controlled datasets and real-world images, and report substantial gains on many real-world debiasing benchmarks (11.4% relative gain on Imagenet-A; 3.2% on BAR, etc). Crucially, we outperform many baselines that incorporate knowledge about known spurious or biased attributes, despite our method not using any such information. We believe that our feature sieve work opens up exciting new research directions in automated adversarial feature extraction & representation learning for deep networks.
|
https://proceedings.mlr.press/v202/tomani23a.html
|
https://proceedings.mlr.press/v202/tomani23a/tomani23a.pdf
|
https://openreview.net/forum?id=taBCtI0m5Y
|
Beyond In-Domain Scenarios: Robust Density-Aware Calibration
|
https://proceedings.mlr.press/v202/tomani23a.html
|
Christian Tomani, Futa Kai Waseda, Yuesong Shen, Daniel Cremers
|
https://proceedings.mlr.press/v202/tomani23a.html
|
ICML 2023
|
Calibrating deep learning models to yield uncertainty-aware predictions is crucial as deep neural networks get increasingly deployed in safety-critical applications. While existing post-hoc calibration methods achieve impressive results on in-domain test datasets, they are limited by their inability to yield reliable uncertainty estimates in domain-shift and out-of-domain (OOD) scenarios. We aim to bridge this gap by proposing DAC, an accuracy-preserving as well as Density-Aware Calibration method based on k-nearest-neighbors (KNN). In contrast to existing post-hoc methods, we utilize hidden layers of classifiers as a source for uncertainty-related information and study their importance. We show that DAC is a generic method that can readily be combined with state-of-the-art post-hoc methods. DAC boosts the robustness of calibration performance in domain-shift and OOD, while maintaining excellent in-domain predictive uncertainty estimates. We demonstrate that DAC leads to consistently better calibration across a large number of model architectures, datasets, and metrics. Additionally, we show that DAC improves calibration substantially on recent large-scale neural networks pre-trained on vast amounts of data.
|
https://proceedings.mlr.press/v202/tong23a.html
|
https://proceedings.mlr.press/v202/tong23a/tong23a.pdf
|
https://openreview.net/forum?id=CERS3hZIrH
|
Distribution Free Domain Generalization
|
https://proceedings.mlr.press/v202/tong23a.html
|
Peifeng Tong, Wu Su, He Li, Jialin Ding, Zhan Haoxiang, Song Xi Chen
|
https://proceedings.mlr.press/v202/tong23a.html
|
ICML 2023
|
Accurate prediction of the out-of-distribution data is desired for a learning algorithm. In domain generalization, training data from source domains tend to have different distributions from that of the target domain, while the target data are absence in the training process. We propose a Distribution Free Domain Generalization (DFDG) procedure for classification by conducting standardization to avoid the dominance of a few domains in the training process. The essence of the DFDG is its reformulating the cross domain/class discrepancy by pairwise two sample test statistics, and equally weights their importance or the covariance structures to avoid dominant domain/class. A theoretical generalization bound is established for the multi-class classification problem. The DFDG is shown to offer a superior performance in empirical studies with fewer hyperparameters, which means faster and easier implementation.
|
https://proceedings.mlr.press/v202/tonin23a.html
|
https://proceedings.mlr.press/v202/tonin23a/tonin23a.pdf
|
https://openreview.net/forum?id=9xqrSeujqc
|
Extending Kernel PCA through Dualization: Sparsity, Robustness and Fast Algorithms
|
https://proceedings.mlr.press/v202/tonin23a.html
|
Francesco Tonin, Alex Lambert, Panagiotis Patrinos, Johan Suykens
|
https://proceedings.mlr.press/v202/tonin23a.html
|
ICML 2023
|
The goal of this paper is to revisit Kernel Principal Component Analysis (KPCA) through dualization of a difference of convex functions. This allows to naturally extend KPCA to multiple objective functions and leads to efficient gradient-based algorithms avoiding the expensive SVD of the Gram matrix. Particularly, we consider objective functions that can be written as Moreau envelopes, demonstrating how to promote robustness and sparsity within the same framework. The proposed method is evaluated on synthetic and realworld benchmarks, showing significant speedup in KPCA training time as well as highlighting the benefits in terms of robustness and sparsity.
|
https://proceedings.mlr.press/v202/tonolini23a.html
|
https://proceedings.mlr.press/v202/tonolini23a/tonolini23a.pdf
|
https://openreview.net/forum?id=YzXkFboejn
|
Robust Weak Supervision with Variational Auto-Encoders
|
https://proceedings.mlr.press/v202/tonolini23a.html
|
Francesco Tonolini, Nikolaos Aletras, Yunlong Jiao, Gabriella Kazai
|
https://proceedings.mlr.press/v202/tonolini23a.html
|
ICML 2023
|
Recent advances in weak supervision (WS) techniques allow to mitigate the enormous cost and effort of human data annotation for supervised machine learning by automating it using simple rule-based labelling functions (LFs). However, LFs need to be carefully designed, often requiring expert domain knowledge and extensive validation for existing WS methods to be effective. To tackle this, we propose the Weak Supervision Variational Auto-Encoder (WS-VAE), a novel framework that combines unsupervised representation learning and weak labelling to reduce the dependence of WS on expert and manual engineering of LFs. Our technique learns from inputs and weak labels jointly to capture the input signals distribution with a latent space. The unsupervised representation component of the WS-VAE regularises the inference of weak labels, while a specifically designed decoder allows the model to learn the relevance of LFs for each input. These unique features lead to considerably improved robustness to the quality of LFs, compared to existing methods. An extensive empirical evaluation on a standard WS benchmark shows that our WS-VAE is competitive to state-of-the-art methods and substantially more robust to LF engineering.
|
https://proceedings.mlr.press/v202/tran23a.html
|
https://proceedings.mlr.press/v202/tran23a/tran23a.pdf
|
https://openreview.net/forum?id=cq5IwsiW5Z
|
Fully Bayesian Autoencoders with Latent Sparse Gaussian Processes
|
https://proceedings.mlr.press/v202/tran23a.html
|
Ba-Hien Tran, Babak Shahbaba, Stephan Mandt, Maurizio Filippone
|
https://proceedings.mlr.press/v202/tran23a.html
|
ICML 2023
|
We present a fully Bayesian autoencoder model that treats both local latent variables and global decoder parameters in a Bayesian fashion. This approach allows for flexible priors and posterior approximations while keeping the inference costs low. To achieve this, we introduce an amortized MCMC approach by utilizing an implicit stochastic network to learn sampling from the posterior over local latent variables. Furthermore, we extend the model by incorporating a Sparse Gaussian Process prior over the latent space, allowing for a fully Bayesian treatment of inducing points and kernel hyperparameters and leading to improved scalability. Additionally, we enable Deep Gaussian Process priors on the latent space and the handling of missing data. We evaluate our model on a range of experiments focusing on dynamic representation learning and generative modeling, demonstrating the strong performance of our approach in comparison to existing methods that combine Gaussian Processes and autoencoders.
|
https://proceedings.mlr.press/v202/trauble23a.html
|
https://proceedings.mlr.press/v202/trauble23a/trauble23a.pdf
|
https://openreview.net/forum?id=LDBIVZCnLl
|
Discrete Key-Value Bottleneck
|
https://proceedings.mlr.press/v202/trauble23a.html
|
Frederik Träuble, Anirudh Goyal, Nasim Rahaman, Michael Curtis Mozer, Kenji Kawaguchi, Yoshua Bengio, Bernhard Schölkopf
|
https://proceedings.mlr.press/v202/trauble23a.html
|
ICML 2023
|
Deep neural networks perform well on classification tasks where data streams are i.i.d. and labeled data is abundant. Challenges emerge with non-stationary training data streams such as continual learning. One powerful approach that has addressed this challenge involves pre-training of large encoders on volumes of readily available data, followed by task-specific tuning. Given a new task, however, updating the weights of these encoders is challenging as a large number of weights needs to be fine-tuned, and as a result, they forget information about the previous tasks. In the present work, we propose a model architecture to address this issue, building upon a discrete bottleneck containing pairs of separate and learnable key-value codes. Our paradigm will be to encode; process the representation via a discrete bottleneck; and decode. Here, the input is fed to the pre-trained encoder, the output of the encoder is used to select the nearest keys, and the corresponding values are fed to the decoder to solve the current task. The model can only fetch and re-use a sparse number of these key-value pairs during inference, enabling localized and context-dependent model updates. We theoretically investigate the ability of the discrete key-value bottleneck to minimize the effect of learning under distribution shifts and show that it reduces the complexity of the hypothesis class. We empirically verify the proposed method under challenging class-incremental learning scenarios and show that the proposed model — without any task boundaries — reduces catastrophic forgetting across a wide variety of pre-trained models, outperforming relevant baselines on this task.
|
https://proceedings.mlr.press/v202/trockman23a.html
|
https://proceedings.mlr.press/v202/trockman23a/trockman23a.pdf
|
https://openreview.net/forum?id=HxN8K1esES
|
Mimetic Initialization of Self-Attention Layers
|
https://proceedings.mlr.press/v202/trockman23a.html
|
Asher Trockman, J Zico Kolter
|
https://proceedings.mlr.press/v202/trockman23a.html
|
ICML 2023
|
It is notoriously difficult to train Transformers on small datasets; typically, large pre-trained models are instead used as the starting point. We explore the weights of such pre-trained Transformers (particularly for vision) to attempt to find reasons for this discrepancy. Surprisingly, we find that simply initializing the weights of self-attention layers so that they "look" more like their pre-trained counterparts allows us to train vanilla Transformers faster and to higher final accuracies, particularly on vision tasks such as CIFAR-10 and ImageNet classification, where we see gains in accuracy of over 5% and 4%, respectively. Our initialization scheme is closed form, learning-free, and very simple: we set the product of the query and key weights to be approximately the identity, and the product of the value and projection weights to approximately the negative identity. As this mimics the patterns we saw in pre-trained Transformers, we call the technique "mimetic initialization".
|
https://proceedings.mlr.press/v202/tsai23a.html
|
https://proceedings.mlr.press/v202/tsai23a/tsai23a.pdf
|
https://openreview.net/forum?id=GLI2hX4vxx
|
Representer Point Selection for Explaining Regularized High-dimensional Models
|
https://proceedings.mlr.press/v202/tsai23a.html
|
Che-Ping Tsai, Jiong Zhang, Hsiang-Fu Yu, Eli Chien, Cho-Jui Hsieh, Pradeep Kumar Ravikumar
|
https://proceedings.mlr.press/v202/tsai23a.html
|
ICML 2023
|
We introduce a novel class of sample-based explanations we term high-dimensional representers, that can be used to explain the predictions of a regularized high-dimensional model in terms of importance weights for each of the training samples. Our workhorse is a novel representer theorem for general regularized high-dimensional models, which decomposes the model prediction in terms of contributions from each of the training samples: with positive (negative) values corresponding to positive (negative) impact training samples to the model’s prediction. We derive consequences for the canonical instances of $\ell_1$ regularized sparse models and nuclear norm regularized low-rank models. As a case study, we further investigate the application of low-rank models in the context of collaborative filtering, where we instantiate high-dimensional representers for specific popular classes of models. Finally, we study the empirical performance of our proposed methods on three real-world binary classification datasets and two recommender system datasets. We also showcase the utility of high-dimensional representers in explaining model recommendations.
|
https://proceedings.mlr.press/v202/tseran23a.html
|
https://proceedings.mlr.press/v202/tseran23a/tseran23a.pdf
|
https://openreview.net/forum?id=mJRMHkaTcp
|
Expected Gradients of Maxout Networks and Consequences to Parameter Initialization
|
https://proceedings.mlr.press/v202/tseran23a.html
|
Hanna Tseran, Guido Montufar
|
https://proceedings.mlr.press/v202/tseran23a.html
|
ICML 2023
|
We study the gradients of a maxout network with respect to inputs and parameters and obtain bounds for the moments depending on the architecture and the parameter distribution. We observe that the distribution of the input-output Jacobian depends on the input, which complicates a stable parameter initialization. Based on the moments of the gradients, we formulate parameter initialization strategies that avoid vanishing and exploding gradients in wide networks. Experiments with deep fully-connected and convolutional networks show that this strategy improves SGD and Adam training of deep maxout networks. In addition, we obtain refined bounds on the expected number of linear regions, results on the expected curve length distortion, and results on the NTK.
|
https://proceedings.mlr.press/v202/tukan23a.html
|
https://proceedings.mlr.press/v202/tukan23a/tukan23a.pdf
|
https://openreview.net/forum?id=QlxwTDQfPp
|
Provable Data Subset Selection For Efficient Neural Networks Training
|
https://proceedings.mlr.press/v202/tukan23a.html
|
Murad Tukan, Samson Zhou, Alaa Maalouf, Daniela Rus, Vladimir Braverman, Dan Feldman
|
https://proceedings.mlr.press/v202/tukan23a.html
|
ICML 2023
|
Radial basis function neural networks (RBFNN) are well-known for their capability to approximate any continuous function on a closed bounded set with arbitrary precision given enough hidden neurons. In this paper, we introduce the first algorithm to construct coresets for RBFNNs, i.e., small weighted subsets that approximate the loss of the input data on any radial basis function network and thus approximate any function defined by an RBFNN on the larger input data. In particular, we construct coresets for radial basis and Laplacian loss functions. We then use our coresets to obtain a provable data subset selection algorithm for training deep neural networks. Since our coresets approximate every function, they also approximate the gradient of each weight in a neural network, which is a particular function on the input. We then perform empirical evaluations on function approximation and dataset subset selection on popular network architectures and data sets, demonstrating the efficacy and accuracy of our coreset construction.
|
https://proceedings.mlr.press/v202/uchendu23a.html
|
https://proceedings.mlr.press/v202/uchendu23a/uchendu23a.pdf
|
https://openreview.net/forum?id=2M7lwN0DTp
|
Jump-Start Reinforcement Learning
|
https://proceedings.mlr.press/v202/uchendu23a.html
|
Ikechukwu Uchendu, Ted Xiao, Yao Lu, Banghua Zhu, Mengyuan Yan, Joséphine Simon, Matthew Bennice, Chuyuan Fu, Cong Ma, Jiantao Jiao, Sergey Levine, Karol Hausman
|
https://proceedings.mlr.press/v202/uchendu23a.html
|
ICML 2023
|
Reinforcement learning (RL) provides a theoretical framework for continuously improving an agent’s behavior via trial and error. However, efficiently learning policies from scratch can be very difficult, particularly for tasks that present exploration challenges. In such settings, it might be desirable to initialize RL with an existing policy, offline data, or demonstrations. However, naively performing such initialization in RL often works poorly, especially for value-based methods. In this paper, we present a meta algorithm that can use offline data, demonstrations, or a pre-existing policy to initialize an RL policy, and is compatible with any RL approach. In particular, we propose Jump-Start Reinforcement Learning (JSRL), an algorithm that employs two policies to solve tasks: a guide-policy, and an exploration-policy. By using the guide-policy to form a curriculum of starting states for the exploration-policy, we are able to efficiently improve performance on a set of simulated robotic tasks. We show via experiments that it is able to significantly outperform existing imitation and reinforcement learning algorithms, particularly in the small-data regime. In addition, we provide an upper bound on the sample complexity of JSRL and show that with the help of a guide-policy, one can improve the sample complexity for non-optimism exploration methods from exponential in horizon to polynomial.
|
https://proceedings.mlr.press/v202/udwani23a.html
|
https://proceedings.mlr.press/v202/udwani23a/udwani23a.pdf
|
https://openreview.net/forum?id=fNMbV07iv7
|
Submodular Order Functions and Assortment Optimization
|
https://proceedings.mlr.press/v202/udwani23a.html
|
Rajan Udwani
|
https://proceedings.mlr.press/v202/udwani23a.html
|
ICML 2023
|
We define a new class of set functions that in addition to being monotone and subadditive, also admit a very limited form of submodularity defined over a permutation of the ground set. We refer to this permutation as a submodular order. We give fast algorithms with strong approximation guarantees for maximizing submodular order functions under a variety of constraints. Applying this new notion to the problem of constrained assortment optimization in fundamental choice models, we obtain new algorithms that are both faster and have stronger approximation guarantees (in some cases, first algorithm with constant factor guarantee). We also show an intriguing connection to the maximization of monotone submodular functions in the streaming model, where we recover best known approximation guarantees as a corollary of our results.
|
https://proceedings.mlr.press/v202/uehara23a.html
|
https://proceedings.mlr.press/v202/uehara23a/uehara23a.pdf
|
https://openreview.net/forum?id=TPQbXPohiz
|
Computationally Efficient PAC RL in POMDPs with Latent Determinism and Conditional Embeddings
|
https://proceedings.mlr.press/v202/uehara23a.html
|
Masatoshi Uehara, Ayush Sekhari, Jason D. Lee, Nathan Kallus, Wen Sun
|
https://proceedings.mlr.press/v202/uehara23a.html
|
ICML 2023
|
We study reinforcement learning with function approximation for large-scale Partially Observable Markov Decision Processes (POMDPs) where the state space and observation space are large or even continuous. Particularly, we consider Hilbert space embeddings of POMDP where the feature of latent states and the feature of observations admit a conditional Hilbert space embedding of the observation emission process, and the latent state transition is deterministic. Under the function approximation setup where the optimal latent state-action $Q$-function is linear in the state feature, and the optimal $Q$-function has a gap in actions, we provide a computationally and statistically efficient algorithm for finding the exact optimal policy. We show our algorithm’s computational and statistical complexities scale polynomially with respect to the horizon and the intrinsic dimension of the feature on the observation space. Furthermore, we show both the deterministic latent transitions and gap assumptions are necessary to avoid statistical complexity exponential in horizon or dimension. Since our guarantee does not have an explicit dependence on the size of the state and observation spaces, our algorithm provably scales to large-scale POMDPs.
|
https://proceedings.mlr.press/v202/ullah23a.html
|
https://proceedings.mlr.press/v202/ullah23a/ullah23a.pdf
|
https://openreview.net/forum?id=IK2mgOCwt3
|
From Adaptive Query Release to Machine Unlearning
|
https://proceedings.mlr.press/v202/ullah23a.html
|
Enayat Ullah, Raman Arora
|
https://proceedings.mlr.press/v202/ullah23a.html
|
ICML 2023
|
We formalize the problem of machine unlearning as design of efficient unlearning algorithms corresponding to learning algorithms which perform a selection of adaptive queries from structured query classes. We give efficient unlearning algorithms for linear and prefix-sum query classes. As applications, we show that unlearning in many problems, in particular, stochastic convex optimization (SCO), can be reduced to the above, yielding improved guarantees for the problem. In particular, for smooth Lipschitz losses and any $\rho>0$, our results yield an unlearning algorithm with excess population risk of $\tilde O\big(\frac{1}{\sqrt{n}}+\frac{\sqrt{d}}{n\rho}\big)$ with unlearning query (gradient) complexity $\tilde O(\rho \cdot \text{Retraining Complexity})$, where $d$ is the model dimensionality and $n$ is the initial number of samples. For non-smooth Lipschitz losses, we give an unlearning algorithm with excess population risk $\tilde O\big(\frac{1}{\sqrt{n}}+\big(\frac{\sqrt{d}}{n\rho}\big)^{1/2}\big)$ with the same unlearning query (gradient) complexity. Furthermore, in the special case of Generalized Linear Models (GLMs), such as those in linear and logistic regression, we get dimension-independent rates of $\tilde O\big(\frac{1}{\sqrt{n}} +\frac{1}{(n\rho)^{2/3}}\big)$ and $\tilde O\big(\frac{1}{\sqrt{n}} +\frac{1}{(n\rho)^{1/3}}\big)$ for smooth Lipschitz and non-smooth Lipschitz losses respectively. Finally, we give generalizations of the above from one unlearning request to dynamic streams consisting of insertions and deletions.
|
https://proceedings.mlr.press/v202/ullah23b.html
|
https://proceedings.mlr.press/v202/ullah23b/ullah23b.pdf
|
https://openreview.net/forum?id=y8qAZhWbNs
|
Private Federated Learning with Autotuned Compression
|
https://proceedings.mlr.press/v202/ullah23b.html
|
Enayat Ullah, Christopher A. Choquette-Choo, Peter Kairouz, Sewoong Oh
|
https://proceedings.mlr.press/v202/ullah23b.html
|
ICML 2023
|
We propose new techniques for reducing communication in private federated learning without the need for setting or tuning compression rates. Our on-the-fly methods automatically adjust the compression rate based on the error induced during training, while maintaining provable privacy guarantees through the use of secure aggregation and differential privacy. Our techniques are provably instance-optimal for mean estimation, meaning that they can adapt to the “hardness of the problem” with minimal interactivity. We demonstrate the effectiveness of our approach on real-world datasets by achieving favorable compression rates without the need for tuning.
|
https://proceedings.mlr.press/v202/uscidda23a.html
|
https://proceedings.mlr.press/v202/uscidda23a/uscidda23a.pdf
|
https://openreview.net/forum?id=U1KcX2JWjF
|
The Monge Gap: A Regularizer to Learn All Transport Maps
|
https://proceedings.mlr.press/v202/uscidda23a.html
|
Théo Uscidda, Marco Cuturi
|
https://proceedings.mlr.press/v202/uscidda23a.html
|
ICML 2023
|
Optimal transport (OT) theory has been used in machine learning to study and characterize maps that can push-forward efficiently a probability measure onto another. Recent works have drawn inspiration from Brenier’s theorem, which states that when the ground cost is the squared-Euclidean distance, the “best” map to morph a continuous measure in $\mathcal{P}(\mathbb{R}^d)$ into another must be the gradient of a convex function. To exploit that result, Makkuva et. al (2020); Korotin et. al (2020) consider maps $T=\nabla f_\theta$, where $f_\theta$ is an input convex neural network (ICNN), as defined by Amos et. al (2017), and fit $\theta$ with SGD using samples. Despite their mathematical elegance, fitting OT maps with ICNNs raises many challenges, due notably to the many constraints imposed on $\theta$; the need to approximate the conjugate of $f_\theta$; or the limitation that they only work for the squared-Euclidean cost. More generally, we question the relevance of using Brenier’s result, which only applies to densities, to constrain the architecture of candidate maps fitted on samples. Motivated by these limitations, we propose a radically different approach to estimating OT maps: Given a cost $c$ and a reference measure $\rho$, we introduce a regularizer, the Monge gap $\mathcal{M}^c_{\rho}(T)$ of a map $T$. That gap quantifies how far a map $T$ deviates from the ideal properties we expect from a $c$-OT map. In practice, we drop all architecture requirements for $T$ and simply minimize a distance (e.g., the Sinkhorn divergence) between $T\sharp\mu$ and $\nu$, regularized by $\mathcal{M}^c_\rho(T)$. We study $\mathcal{M}^c_{\rho}$ and show how our simple pipeline significantly outperforms other baselines in practice.
|
https://proceedings.mlr.press/v202/vacher23a.html
|
https://proceedings.mlr.press/v202/vacher23a/vacher23a.pdf
|
https://openreview.net/forum?id=zodnF0pqK7
|
Semi-Dual Unbalanced Quadratic Optimal Transport: fast statistical rates and convergent algorithm.
|
https://proceedings.mlr.press/v202/vacher23a.html
|
Adrien Vacher, François-Xavier Vialard
|
https://proceedings.mlr.press/v202/vacher23a.html
|
ICML 2023
|
In this paper, we derive a semi-dual formulation for the problem of unbalanced quadratic optimal transport and we study its stability properties, namely we give upper and lower bounds for the Bregman divergence of the new objective that hold globally. We observe that the new objective gains even more convexity than in the balanced case. We use this formulation to prove the first results on statistical estimation of UOT potentials and we leverage the extra convexity to recover super-parametric rates. Interestingly, unlike in the balanced case, we do not require the potentials to be smooth. Then, use variable metric descent to solve the semi-dual problem for which we prove convergence at a $1/k$ rate for strongly convex potentials and exponential convergence in the balanced case when potentials are also smooth. We emphasize that our convergence results has an interest on its own as it generalizes previous convergence results to non-equivalent metrics. Last, we instantiate a proof-of-concept tractable version of our theoretical algorithm that we benchmark on a 2D experiment in the balanced case and on a medium dimension synthetic experiment in the unbalanced case.
|
https://proceedings.mlr.press/v202/vadeboncoeur23a.html
|
https://proceedings.mlr.press/v202/vadeboncoeur23a/vadeboncoeur23a.pdf
|
https://openreview.net/forum?id=g6WlWFFZxa
|
Random Grid Neural Processes for Parametric Partial Differential Equations
|
https://proceedings.mlr.press/v202/vadeboncoeur23a.html
|
Arnaud Vadeboncoeur, Ieva Kazlauskaite, Yanni Papandreou, Fehmi Cirak, Mark Girolami, Omer Deniz Akyildiz
|
https://proceedings.mlr.press/v202/vadeboncoeur23a.html
|
ICML 2023
|
We introduce a new class of spatially stochastic physics and data informed deep latent models for parametric partial differential equations (PDEs) which operate through scalable variational neural processes. We achieve this by assigning probability measures to the spatial domain, which allows us to treat collocation grids probabilistically as random variables to be marginalised out. Adapting this spatial statistics view, we solve forward and inverse problems for parametric PDEs in a way that leads to the construction of Gaussian process models of solution fields. The implementation of these random grids poses a unique set of challenges for inverse physics informed deep learning frameworks and we propose a new architecture called Grid Invariant Convolutional Networks (GICNets) to overcome these challenges. We further show how to incorporate noisy data in a principled manner into our physics informed model to improve predictions for problems where data may be available but whose measurement location does not coincide with any fixed mesh or grid. The proposed method is tested on a nonlinear Poisson problem, Burgers equation, and Navier-Stokes equations, and we provide extensive numerical comparisons. We demonstrate significant computational advantages over current physics informed neural learning methods for parametric PDEs while improving the predictive capabilities and flexibility of these models.
|
https://proceedings.mlr.press/v202/vakili23a.html
|
https://proceedings.mlr.press/v202/vakili23a/vakili23a.pdf
|
https://openreview.net/forum?id=RNRbovY8zV
|
Delayed Feedback in Kernel Bandits
|
https://proceedings.mlr.press/v202/vakili23a.html
|
Sattar Vakili, Danyal Ahmed, Alberto Bernacchia, Ciara Pike-Burke
|
https://proceedings.mlr.press/v202/vakili23a.html
|
ICML 2023
|
Black box optimisation of an unknown function from expensive and noisy evaluations is a ubiquitous problem in machine learning, academic research and industrial production. An abstraction of the problem can be formulated as a kernel based bandit problem (also known as Bayesian optimisation), where a learner aims at optimising a kernelized function through sequential noisy observations. The existing work predominantly assumes feedback is immediately available; an assumption which fails in many real world situations, including recommendation systems, clinical trials and hyperparameter tuning. We consider a kernel bandit problem under stochastically delayed feedback, and propose an algorithm with $\tilde{\mathcal{O}}\left(\sqrt{\Gamma_k(T) T}+\mathbb{E}[\tau]\right)$ regret, where $T$ is the number of time steps, $\Gamma_k(T)$ is the maximum information gain of the kernel with $T$ observations, and $\tau$ is the delay random variable. This represents a significant improvement over the state of the art regret bound of $\tilde{\mathcal{O}}\left(\Gamma_k(T)\sqrt{ T}+\mathbb{E}[\tau]\Gamma_k(T)\right)$ reported in (Verma et al., 2022). In particular, for very non-smooth kernels, the information gain grows almost linearly in time, trivializing the existing results. We also validate our theoretical results with simulations.
|
https://proceedings.mlr.press/v202/van-breugel23a.html
|
https://proceedings.mlr.press/v202/van-breugel23a/van-breugel23a.pdf
|
https://openreview.net/forum?id=I5kywOUcl7
|
Synthetic Data, Real Errors: How (Not) to Publish and Use Synthetic Data
|
https://proceedings.mlr.press/v202/van-breugel23a.html
|
Boris Van Breugel, Zhaozhi Qian, Mihaela Van Der Schaar
|
https://proceedings.mlr.press/v202/van-breugel23a.html
|
ICML 2023
|
Generating synthetic data through generative models is gaining interest in the ML community and beyond, promising a future where datasets can be tailored to individual needs. Unfortunately, synthetic data is usually not perfect, resulting in potential errors in downstream tasks. In this work we explore how the generative process affects the downstream ML task. We show that the naive synthetic data approach—using synthetic data as if it is real—leads to downstream models and analyses that do not generalize well to real data. As a first step towards better ML in the synthetic data regime, we introduce Deep Generative Ensemble (DGE)—a framework inspired by Deep Ensembles that aims to implicitly approximate the posterior distribution over the generative process model parameters. DGE improves downstream model training, evaluation, and uncertainty quantification, vastly outperforming the naive approach on average. The largest improvements are achieved for minority classes and low-density regions of the original data, for which the generative uncertainty is largest.
|
https://proceedings.mlr.press/v202/van-der-hoeven23a.html
|
https://proceedings.mlr.press/v202/van-der-hoeven23a/van-der-hoeven23a.pdf
|
https://openreview.net/forum?id=zIVu5Yidhm
|
Trading-Off Payments and Accuracy in Online Classification with Paid Stochastic Experts
|
https://proceedings.mlr.press/v202/van-der-hoeven23a.html
|
Dirk Van Der Hoeven, Ciara Pike-Burke, Hao Qiu, Nicolò Cesa-Bianchi
|
https://proceedings.mlr.press/v202/van-der-hoeven23a.html
|
ICML 2023
|
We investigate online classification with paid stochastic experts. Here, before making their prediction, each expert must be paid. The amount that we pay each expert directly influences the accuracy of their prediction through some unknown Lipschitz “productivity” function. In each round, the learner must decide how much to pay each expert and then make a prediction. They incur a cost equal to a weighted sum of the prediction error and upfront payments for all experts. We introduce an online learning algorithm whose total cost after $T$ rounds exceeds that of a predictor which knows the productivity of all experts in advance by at most $\mathcal{O}\big(K^2(\ln T)\sqrt{T}\big)$ where $K$ is the number of experts. In order to achieve this result, we combine Lipschitz bandits and online classification with surrogate losses. These tools allow us to improve upon the bound of order $T^{2/3}$ one would obtain in the standard Lipschitz bandit setting. Our algorithm is empirically evaluated on synthetic data.
|
https://proceedings.mlr.press/v202/van-der-laan23a.html
|
https://proceedings.mlr.press/v202/van-der-laan23a/van-der-laan23a.pdf
|
https://openreview.net/forum?id=nuHWrVbmus
|
Causal Isotonic Calibration for Heterogeneous Treatment Effects
|
https://proceedings.mlr.press/v202/van-der-laan23a.html
|
Lars Van Der Laan, Ernesto Ulloa-Perez, Marco Carone, Alex Luedtke
|
https://proceedings.mlr.press/v202/van-der-laan23a.html
|
ICML 2023
|
We propose causal isotonic calibration, a novel nonparametric method for calibrating predictors of heterogeneous treatment effects. Furthermore, we introduce cross-calibration, a data-efficient variant of calibration that eliminates the need for hold-out calibration sets. Cross-calibration leverages cross-fitted predictors and generates a single calibrated predictor using all available data. Under weak conditions that do not assume monotonicity, we establish that both causal isotonic calibration and cross-calibration achieve fast doubly-robust calibration rates, as long as either the propensity score or outcome regression is estimated accurately in a suitable sense. The proposed causal isotonic calibrator can be wrapped around any black-box learning algorithm, providing robust and distribution-free calibration guarantees while preserving predictive performance.
|
https://proceedings.mlr.press/v202/vanderschueren23a.html
|
https://proceedings.mlr.press/v202/vanderschueren23a/vanderschueren23a.pdf
|
https://openreview.net/forum?id=qaPmiAGZcV
|
Accounting For Informative Sampling When Learning to Forecast Treatment Outcomes Over Time
|
https://proceedings.mlr.press/v202/vanderschueren23a.html
|
Toon Vanderschueren, Alicia Curth, Wouter Verbeke, Mihaela Van Der Schaar
|
https://proceedings.mlr.press/v202/vanderschueren23a.html
|
ICML 2023
|
Machine learning (ML) holds great potential for accurately forecasting treatment outcomes over time, which could ultimately enable the adoption of more individualized treatment strategies in many practical applications. However, a significant challenge that has been largely overlooked by the ML literature on this topic is the presence of informative sampling in observational data. When instances are observed irregularly over time, sampling times are typically not random, but rather informative–depending on the instance’s characteristics, past outcomes, and administered treatments. In this work, we formalize informative sampling as a covariate shift problem and show that it can prohibit accurate estimation of treatment outcomes if not properly accounted for. To overcome this challenge, we present a general framework for learning treatment outcomes in the presence of informative sampling using inverse intensity-weighting, and propose a novel method, TESAR-CDE, that instantiates this framework using Neural CDEs. Using a simulation environment based on a clinical use case, we demonstrate the effectiveness of our approach in learning under informative sampling.
|
https://proceedings.mlr.press/v202/vannella23a.html
|
https://proceedings.mlr.press/v202/vannella23a/vannella23a.pdf
|
https://openreview.net/forum?id=AzFq5HxVlg
|
Best Arm Identification in Multi-Agent Multi-Armed Bandits
|
https://proceedings.mlr.press/v202/vannella23a.html
|
Filippo Vannella, Alexandre Proutiere, Jaeseong Jeong
|
https://proceedings.mlr.press/v202/vannella23a.html
|
ICML 2023
|
We investigate the problem of best arm identification in Multi-Agent Multi-Armed Bandits (MAMABs) where the rewards are defined through a factor graph. The objective is to find an optimal global action with a prescribed level of confidence and minimal sample complexity. We derive a tight instance-specific lower bound of the sample complexity and characterize the corresponding optimal sampling strategy. Unfortunately, this bound is obtained by solving a combinatorial optimization problem with a number of variables and constraints exponentially growing with the number of agents. We leverage Mean Field (MF) techniques to obtain, in a computationally efficient manner, an approximation of the lower bound. The approximation scales at most as $\rho K^d$ (where $\rho$, $K$, and $d$ denote the number of factors in the graph, the number of possible actions per agent, and the maximal degree of the factor graph). We devise MF-TaS (Mean-Field-Track-and-Stop), an algorithm whose sample complexity provably matches our approximated lower bound. We illustrate the performance of MF-TaS numerically using both synthetic and real-world experiments (e.g., to solve the antenna tilt optimization problem in radio communication networks).
|
https://proceedings.mlr.press/v202/varma23a.html
|
https://proceedings.mlr.press/v202/varma23a/varma23a.pdf
|
https://openreview.net/forum?id=R7X1sTaM6J
|
Conditional Tree Matching for Inference-Time Adaptation of Tree Prediction Models
|
https://proceedings.mlr.press/v202/varma23a.html
|
Harshit Varma, Abhijeet Awasthi, Sunita Sarawagi
|
https://proceedings.mlr.press/v202/varma23a.html
|
ICML 2023
|
We present CTreeOT, a convergent, differentiable algorithm for matching two trees when each tree is conditioned on some input. Such conditional tree matching is useful for light-weight, few-shot adaptation of tree prediction models without parameter fine-tuning. CTreeOT includes an alignment algorithm that extends the popular Sinkhorn algorithm for matching tree nodes while supporting constraints on tree edges. The algorithm involves alternating between matrix rescaling and message passing updates, and can be efficiently expressed as GPU tensor operations. The second part of CTreeOT is fine-grained relevance-based reweighting of nodes that makes the match scores useful for prediction tasks. We demonstrate the usefulness of CTreeOT for cross-schema adaptation of Text-to-SQL, a popular semantic parsing task. We show that compared to state-of-the-art methods, we achieve significant increase in adaptation accuracy.
|
https://proceedings.mlr.press/v202/veldt23a.html
|
https://proceedings.mlr.press/v202/veldt23a/veldt23a.pdf
|
https://openreview.net/forum?id=eZK32L3Pl2
|
Optimal LP Rounding and Linear-Time Approximation Algorithms for Clustering Edge-Colored Hypergraphs
|
https://proceedings.mlr.press/v202/veldt23a.html
|
Nate Veldt
|
https://proceedings.mlr.press/v202/veldt23a.html
|
ICML 2023
|
We study the approximability of an existing framework for clustering edge-colored hypergraphs, which is closely related to chromatic correlation clustering and is motivated by machine learning and data mining applications where the goal is to cluster a set of objects based on multiway interactions of different categories or types. We present improved approximation guarantees based on linear programming, and show they are tight by proving a matching integrality gap. Our results also include new approximation hardness results, a combinatorial 2-approximation whose runtime is linear in the hypergraph size, and several new connections to well-studied objectives such as vertex cover and hypergraph multiway cut.
|
https://proceedings.mlr.press/v202/velingker23a.html
|
https://proceedings.mlr.press/v202/velingker23a/velingker23a.pdf
|
https://openreview.net/forum?id=Iey50XHA3g
|
Fast $(1+\varepsilon)$-Approximation Algorithms for Binary Matrix Factorization
|
https://proceedings.mlr.press/v202/velingker23a.html
|
Ameya Velingker, Maximilian Vötsch, David Woodruff, Samson Zhou
|
https://proceedings.mlr.press/v202/velingker23a.html
|
ICML 2023
|
We introduce efficient $(1+\varepsilon)$-approximation algorithms for the binary matrix factorization (BMF) problem, where the inputs are a matrix $\mathbf{A}\in\{0,1\}^{n\times d}$, a rank parameter $k>0$, as well as an accuracy parameter $\varepsilon>0$, and the goal is to approximate $\mathbf{A}$ as a product of low-rank factors $\mathbf{U}\in\{0,1\}^{n\times k}$ and $\mathbf{V}\in\{0,1\}^{k\times d}$. Equivalently, we want to find $\mathbf{U}$ and $\mathbf{V}$ that minimize the Frobenius loss $\|\mathbf{U}\mathbf{V} - \mathbf{A}\|_F^2$. Before this work, the state-of-the-art for this problem was the approximation algorithm of Kumar et. al. [ICML 2019], which achieves a $C$-approximation for some constant $C\ge 576$. We give the first $(1+\varepsilon)$-approximation algorithm using running time singly exponential in $k$, where $k$ is typically a small integer. Our techniques generalize to other common variants of the BMF problem, admitting bicriteria $(1+\varepsilon)$-approximation algorithms for $L_p$ loss functions and the setting where matrix operations are performed in $\mathbb{F}_2$. Our approach can be implemented in standard big data models, such as the streaming or distributed models.
|
https://proceedings.mlr.press/v202/vemula23a.html
|
https://proceedings.mlr.press/v202/vemula23a/vemula23a.pdf
|
https://openreview.net/forum?id=F6uHGKVa04
|
The Virtues of Laziness in Model-based RL: A Unified Objective and Algorithms
|
https://proceedings.mlr.press/v202/vemula23a.html
|
Anirudh Vemula, Yuda Song, Aarti Singh, Drew Bagnell, Sanjiban Choudhury
|
https://proceedings.mlr.press/v202/vemula23a.html
|
ICML 2023
|
We propose a novel approach to addressing two fundamental challenges in Model-based Reinforcement Learning (MBRL): the computational expense of repeatedly finding a good policy in the learned model, and the objective mismatch between model fitting and policy computation. Our "lazy" method leverages a novel unified objective, Performance Difference via Advantage in Model, to capture the performance difference between the learned policy and expert policy under the true dynamics. This objective demonstrates that optimizing the expected policy advantage in the learned model under an exploration distribution is sufficient for policy computation, resulting in a significant boost in computational efficiency compared to traditional planning methods. Additionally, the unified objective uses a value moment matching term for model fitting, which is aligned with the model’s usage during policy computation. We present two no-regret algorithms to optimize the proposed objective, and demonstrate their statistical and computational gains compared to existing MBRL methods through simulated benchmarks.
|
https://proceedings.mlr.press/v202/venturini23a.html
|
https://proceedings.mlr.press/v202/venturini23a/venturini23a.pdf
|
https://openreview.net/forum?id=X8h8wLjog7
|
Learning the Right Layers a Data-Driven Layer-Aggregation Strategy for Semi-Supervised Learning on Multilayer Graphs
|
https://proceedings.mlr.press/v202/venturini23a.html
|
Sara Venturini, Andrea Cristofari, Francesco Rinaldi, Francesco Tudisco
|
https://proceedings.mlr.press/v202/venturini23a.html
|
ICML 2023
|
Clustering (or community detection) on multilayer graphs poses several additional complications with respect to standard graphs as different layers may be characterized by different structures and types of information. One of the major challenges is to establish the extent to which each layer contributes to the cluster assignment in order to effectively take advantage of the multilayer structure and improve upon the classification obtained using the individual layers or their union. However, making an informed a-priori assessment about the clustering information content of the layers can be very complicated. In this work, we assume a semi-supervised learning setting, where the class of a small percentage of nodes is initially provided, and we propose a parameter-free Laplacian-regularized model that learns an optimal nonlinear combination of the different layers from the available input labels. The learning algorithm is based on a Frank-Wolfe optimization scheme with inexact gradient, combined with a modified Label Propagation iteration. We provide a detailed convergence analysis of the algorithm and extensive experiments on synthetic and real-world datasets, showing that the proposed method compares favourably with a variety of baselines and outperforms each individual layer when used in isolation.
|
https://proceedings.mlr.press/v202/venuto23a.html
|
https://proceedings.mlr.press/v202/venuto23a/venuto23a.pdf
|
https://openreview.net/forum?id=5RvZb6Lcbz
|
Multi-Environment Pretraining Enables Transfer to Action Limited Datasets
|
https://proceedings.mlr.press/v202/venuto23a.html
|
David Venuto, Sherry Yang, Pieter Abbeel, Doina Precup, Igor Mordatch, Ofir Nachum
|
https://proceedings.mlr.press/v202/venuto23a.html
|
ICML 2023
|
Using massive datasets to train large-scale models has emerged as a dominant approach for broad generalization in natural language and vision applications. In reinforcement learning, however, a key challenge is that available data of sequential decision making is often not annotated with actions - for example, videos of game-play are much more available than sequences of frames paired with their logged game controls. We propose to circumvent this challenge by combining large but sparsely-annotated datasets from a target environment of interest with fully-annotated datasets from various other source environments. Our method, Action Limited PreTraining (ALPT), leverages the generalization capabilities of inverse dynamics modelling (IDM) to label missing action data in the target environment. We show that utilizing even one additional environment dataset of labelled data during IDM pretraining gives rise to substantial improvements in generating action labels for unannotated sequences. We evaluate our method on benchmark game-playing environments and show that we can significantly improve game performance and generalization capability compared to other approaches, using annotated datasets equivalent to only $12$ minutes of gameplay. Highlighting the power of IDM, we show that these benefits remain even when target and source environments share no common actions.
|
https://proceedings.mlr.press/v202/verma23a.html
|
https://proceedings.mlr.press/v202/verma23a/verma23a.pdf
|
https://openreview.net/forum?id=EB5unD2ojL
|
AbODE: Ab initio antibody design using conjoined ODEs
|
https://proceedings.mlr.press/v202/verma23a.html
|
Yogesh Verma, Markus Heinonen, Vikas Garg
|
https://proceedings.mlr.press/v202/verma23a.html
|
ICML 2023
|
Antibodies are Y-shaped proteins that neutralize pathogens and constitute the core of our adaptive immune system. De novo generation of new antibodies that target specific antigens holds the key to accelerating vaccine discovery. However, this co-design of the amino acid sequence and the 3D structure subsumes and accentuates, some central challenges from multiple tasks including protein folding (sequence to structure), inverse folding (structure to sequence), and docking (binding). We strive to surmount these challenges with a new generative model AbODE that extends graph PDEs to accommodate both contextual information and external interactions. Unlike existing approaches, AbODE uses a single round of full-shot decoding, and elicits continuous differential attention that encapsulates, and evolves with, latent interactions within the antibody as well as those involving the antigen. We unravel fundamental connections between AbODE and temporal networks as well as graph-matching networks. The proposed model significantly outperforms existing methods on standard metrics across benchmarks.
|
https://proceedings.mlr.press/v202/vero23a.html
|
https://proceedings.mlr.press/v202/vero23a/vero23a.pdf
|
https://openreview.net/forum?id=mRiDy4qGwB
|
TabLeak: Tabular Data Leakage in Federated Learning
|
https://proceedings.mlr.press/v202/vero23a.html
|
Mark Vero, Mislav Balunovic, Dimitar Iliev Dimitrov, Martin Vechev
|
https://proceedings.mlr.press/v202/vero23a.html
|
ICML 2023
|
While federated learning (FL) promises to preserve privacy, recent works in the image and text domains have shown that training updates leak private client data. However, most high-stakes applications of FL (e.g., in healthcare and finance) use tabular data, where the risk of data leakage has not yet been explored. A successful attack for tabular data must address two key challenges unique to the domain: (i) obtaining a solution to a high-variance mixed discrete-continuous optimization problem, and (ii) enabling human assessment of the reconstruction as unlike for image and text data, direct human inspection is not possible. In this work we address these challenges and propose TabLeak, the first comprehensive reconstruction attack on tabular data. TabLeak is based on two key contributions: (i) a method which leverages a softmax relaxation and pooled ensembling to solve the optimization problem, and (ii) an entropy-based uncertainty quantification scheme to enable human assessment. We evaluate TabLeak on four tabular datasets for both FedSGD and FedAvg training protocols, and show that it successfully breaks several settings previously deemed safe. For instance, we extract large subsets of private data at $>$90% accuracy even at the large batch size of 128. Our findings demonstrate that current high-stakes tabular FL is excessively vulnerable to leakage attacks.
|
https://proceedings.mlr.press/v202/vicol23a.html
|
https://proceedings.mlr.press/v202/vicol23a/vicol23a.pdf
|
https://openreview.net/forum?id=K0InBsKODr
|
Low-Variance Gradient Estimation in Unrolled Computation Graphs with ES-Single
|
https://proceedings.mlr.press/v202/vicol23a.html
|
Paul Vicol
|
https://proceedings.mlr.press/v202/vicol23a.html
|
ICML 2023
|
We propose an evolution strategies-based algorithm for estimating gradients in unrolled computation graphs, called ES-Single. Similarly to the recently-proposed Persistent Evolution Strategies (PES), ES-Single is unbiased, and overcomes chaos arising from recursive function applications by smoothing the meta-loss landscape. ES-Single samples a single perturbation per particle, that is kept fixed over the course of an inner problem (e.g., perturbations are not re-sampled for each partial unroll). Compared to PES, ES-Single is simpler to implement and has lower variance: the variance of ES-Single is constant with respect to the number of truncated unrolls, removing a key barrier in applying ES to long inner problems using short truncations. We show that ES-Single is unbiased for quadratic inner problems, and demonstrate empirically that its variance can be substantially lower than that of PES. ES-Single consistently outperforms PES on a variety of tasks, including a synthetic benchmark task, hyperparameter optimization, training recurrent neural networks, and training learned optimizers.
|
https://proceedings.mlr.press/v202/vilnis23a.html
|
https://proceedings.mlr.press/v202/vilnis23a/vilnis23a.pdf
|
https://openreview.net/forum?id=EfhmBBrXY2
|
Arithmetic Sampling: Parallel Diverse Decoding for Large Language Models
|
https://proceedings.mlr.press/v202/vilnis23a.html
|
Luke Vilnis, Yury Zemlyanskiy, Patrick Murray, Alexandre Tachard Passos, Sumit Sanghai
|
https://proceedings.mlr.press/v202/vilnis23a.html
|
ICML 2023
|
Decoding methods for large language models often trade-off between diversity of outputs and parallelism of computation. Methods such as beam search and Gumbel top-k sampling can guarantee a different output for each element of the beam, but are not easy to parallelize. Alternatively, methods such as temperature sampling and its modifications (top-k sampling, nucleus sampling, typical decoding, and others), are embarrassingly parallel, but have no guarantees about duplicate samples. We present a framework for sampling according to an arithmetic code book implicitly defined by a large language model, compatible with common sampling variations, with provable beam diversity under certain conditions, as well as being embarrassingly parallel and providing unbiased and consistent expectations from the original model. We demonstrate the effectiveness of our approach on WMT machine translation, more than halving the standard deviation when estimating expected BLEU score reward, and closing the BLEU score gap between independent sampling and beam search by up to 63%.
|
https://proceedings.mlr.press/v202/voloshin23a.html
|
https://proceedings.mlr.press/v202/voloshin23a/voloshin23a.pdf
|
https://openreview.net/forum?id=wCy3pef6kA
|
Eventual Discounting Temporal Logic Counterfactual Experience Replay
|
https://proceedings.mlr.press/v202/voloshin23a.html
|
Cameron Voloshin, Abhinav Verma, Yisong Yue
|
https://proceedings.mlr.press/v202/voloshin23a.html
|
ICML 2023
|
Linear temporal logic (LTL) offers a simplified way of specifying tasks for policy optimization that may otherwise be difficult to describe with scalar reward functions. However, the standard RL framework can be too myopic to find maximally LTL satisfying policies. This paper makes two contributions. First, we develop a new value-function based proxy, using a technique we call eventual discounting, under which one can find policies that satisfy the LTL specification with highest achievable probability. Second, we develop a new experience replay method for generating off-policy data from on-policy rollouts via counterfactual reasoning on different ways of satisfying the LTL specification. Our experiments, conducted in both discrete and continuous state-action spaces, confirm the effectiveness of our counterfactual experience replay approach.
|
https://proceedings.mlr.press/v202/von-oswald23a.html
|
https://proceedings.mlr.press/v202/von-oswald23a/von-oswald23a.pdf
|
https://openreview.net/forum?id=tHvXrFQma5
|
Transformers Learn In-Context by Gradient Descent
|
https://proceedings.mlr.press/v202/von-oswald23a.html
|
Johannes Von Oswald, Eyvind Niklasson, Ettore Randazzo, Joao Sacramento, Alexander Mordvintsev, Andrey Zhmoginov, Max Vladymyrov
|
https://proceedings.mlr.press/v202/von-oswald23a.html
|
ICML 2023
|
At present, the mechanisms of in-context learning in Transformers are not well understood and remain mostly an intuition. In this paper, we suggest that training Transformers on auto-regressive objectives is closely related to gradient-based meta-learning formulations. We start by providing a simple weight construction that shows the equivalence of data transformations induced by 1) a single linear self-attention layer and by 2) gradient-descent (GD) on a regression loss. Motivated by that construction, we show empirically that when training self-attention-only Transformers on simple regression tasks either the models learned by GD and Transformers show great similarity or, remarkably, the weights found by optimization match the construction. Thus we show how trained Transformers become mesa-optimizers i.e. learn models by gradient descent in their forward pass. This allows us, at least in the domain of regression problems, to mechanistically understand the inner workings of in-context learning in optimized Transformers. Building on this insight, we furthermore identify how Transformers surpass the performance of plain gradient descent by learning an iterative curvature correction and learn linear models on deep data representations to solve non-linear regression tasks. Finally, we discuss intriguing parallels to a mechanism identified to be crucial for in-context learning termed induction-head (Olsson et al., 2022) and show how it could be understood as a specific case of in-context learning by gradient descent learning within Transformers.
|
https://proceedings.mlr.press/v202/von-rohrscheidt23a.html
|
https://proceedings.mlr.press/v202/von-rohrscheidt23a/von-rohrscheidt23a.pdf
|
https://openreview.net/forum?id=c6Wg91Xpbe
|
Topological Singularity Detection at Multiple Scales
|
https://proceedings.mlr.press/v202/von-rohrscheidt23a.html
|
Julius Von Rohrscheidt, Bastian Rieck
|
https://proceedings.mlr.press/v202/von-rohrscheidt23a.html
|
ICML 2023
|
The manifold hypothesis, which assumes that data lies on or close to an unknown manifold of low intrinsic dimension, is a staple of modern machine learning research. However, recent work has shown that real-world data exhibits distinct non-manifold structures, i.e. singularities, that can lead to erroneous findings. Detecting such singularities is therefore crucial as a precursor to interpolation and inference tasks. We address this issue by developing a topological framework that (i) quantifies the local intrinsic dimension, and (ii) yields a Euclidicity score for assessing the ’manifoldness’ of a point along multiple scales. Our approach identifies singularities of complex spaces, while also capturing singular structures and local geometric complexity in image data.
|
https://proceedings.mlr.press/v202/voracek23a.html
|
https://proceedings.mlr.press/v202/voracek23a/voracek23a.pdf
|
https://openreview.net/forum?id=vPLIRidmYO
|
Improving l1-Certified Robustness via Randomized Smoothing by Leveraging Box Constraints
|
https://proceedings.mlr.press/v202/voracek23a.html
|
Vaclav Voracek, Matthias Hein
|
https://proceedings.mlr.press/v202/voracek23a.html
|
ICML 2023
|
Randomized smoothing is a popular method to certify robustness of image classifiers to adversarial input perturbations. It is the only certification technique which scales directly to datasets of higher dimension such as ImageNet. However, current techniques are not able to utilize the fact that any adversarial example has to lie in the image space, that is $[0,1]^d$; otherwise, one can trivially detect it. To address this suboptimality, we derive new certification formulae which lead to significant improvements in the certified $\ell_1$-robustness without the need of adapting the classifiers or change of smoothing distributions. The code is released at https://github.com/vvoracek/L1-smoothing
|
https://proceedings.mlr.press/v202/vuong23a.html
|
https://proceedings.mlr.press/v202/vuong23a/vuong23a.pdf
|
https://openreview.net/forum?id=eh4403nqwh
|
Vector Quantized Wasserstein Auto-Encoder
|
https://proceedings.mlr.press/v202/vuong23a.html
|
Long Tung Vuong, Trung Le, He Zhao, Chuanxia Zheng, Mehrtash Harandi, Jianfei Cai, Dinh Phung
|
https://proceedings.mlr.press/v202/vuong23a.html
|
ICML 2023
|
Learning deep discrete latent presentations offers a promise of better symbolic and summarized abstractions that are more useful to subsequent downstream tasks. Inspired by the seminal Vector Quantized Variational Auto-Encoder (VQ-VAE), most of work in learning deep discrete representations has mainly focused on improving the original VQ-VAE form and none of them has studied learning deep discrete representations from the generative viewpoint. In this work, we study learning deep discrete representations from the generative viewpoint. Specifically, we endow discrete distributions over sequences of codewords and learn a deterministic decoder that transports the distribution over the sequences of codewords to the data distribution via minimizing a WS distance between them. We develop further theories to connect it with the clustering viewpoint of WS distance, allowing us to have a better and more controllable clustering solution. Finally, we empirically evaluate our method on several well-known benchmarks, where it achieves better qualitative and quantitative performances than the other VQ-VAE variants in terms of the codebook utilization and image reconstruction/generation.
|
https://proceedings.mlr.press/v202/vyas23a.html
|
https://proceedings.mlr.press/v202/vyas23a/vyas23a.pdf
|
https://openreview.net/forum?id=HeuWdhGNk4
|
Competitive Gradient Optimization
|
https://proceedings.mlr.press/v202/vyas23a.html
|
Abhijeet Vyas, Brian Bullins, Kamyar Azizzadenesheli
|
https://proceedings.mlr.press/v202/vyas23a.html
|
ICML 2023
|
We study the problem of convergence to a stationary point in zero-sum games. We propose competitive gradient optimization (CGO), a gradient-based method that incorporates the interactions between two players in zero-sum games for its iterative updates. We provide a continuous-time analysis of CGO and its convergence properties while showing that in the continuous limit, previous methods degenerate to their gradient descent ascent (GDA) variants. We further provide a rate of convergence to stationary points in the discrete-time setting. We propose a generalized class of $\alpha$-coherent functions and show that for strictly $\alpha$-coherent functions, CGO ensures convergence to a saddle point. Moreover, we propose optimistic CGO (oCGO), an optimistic variant, for which we show a convergence rate of $O(\frac{1}{n})$ to saddle points for $\alpha$-coherent functions.
|
https://proceedings.mlr.press/v202/vyas23b.html
|
https://proceedings.mlr.press/v202/vyas23b/vyas23b.pdf
|
https://openreview.net/forum?id=qRAHZVnQNY
|
On Provable Copyright Protection for Generative Models
|
https://proceedings.mlr.press/v202/vyas23b.html
|
Nikhil Vyas, Sham M. Kakade, Boaz Barak
|
https://proceedings.mlr.press/v202/vyas23b.html
|
ICML 2023
|
There is a growing concern that learned conditional generative models may output samples that are substantially similar to some copyrighted data $C$ that was in their training set. We give a formal definition of near access-freeness (NAF) and prove bounds on the probability that a model satisfying this definition outputs a sample similar to $C$, even if $C$ is included in its training set. Roughly speaking, a generative model $p$ is $k$-NAF if for every potentially copyrighted data $C$, the output of $p$ diverges by at most $k$-bits from the output of a model $q$ that did not access $C$ at all. We also give generative model learning algorithms, which efficiently modify the original generative model learning algorithm in a black box manner, that output generative models with strong bounds on the probability of sampling protected content. Furthermore, we provide promising experiments for both language (transformers) and image (diffusion) generative models, showing minimal degradation in output quality while ensuring strong protections against sampling protected content.
|
https://proceedings.mlr.press/v202/wagenmaker23a.html
|
https://proceedings.mlr.press/v202/wagenmaker23a/wagenmaker23a.pdf
|
https://openreview.net/forum?id=hFcIR2tUUi
|
Leveraging Offline Data in Online Reinforcement Learning
|
https://proceedings.mlr.press/v202/wagenmaker23a.html
|
Andrew Wagenmaker, Aldo Pacchiano
|
https://proceedings.mlr.press/v202/wagenmaker23a.html
|
ICML 2023
|
Two central paradigms have emerged in the reinforcement learning (RL) community: online RL and offline RL. In the online RL setting, the agent has no prior knowledge of the environment, and must interact with it in order to find an $\epsilon$-optimal policy. In the offline RL setting, the learner instead has access to a fixed dataset to learn from, but is unable to otherwise interact with the environment, and must obtain the best policy it can from this offline data. Practical scenarios often motivate an intermediate setting: if we have some set of offline data and may also interact with the environment, how can we best use the offline data to minimize the number of online interactions necessary to learn an $\epsilon$-optimal policy. In this work, we consider this setting, which we call the FineTuneRL setting, for MDPs with linear structure. We characterize the necessary number of online samples needed in this setting given access to some offline dataset, and develop an algorithm, FTPedel, which is provably optimal, up to $H$ factors. We show through an explicit example that combining offline data with online interactions can lead to a provable improvement over either purely offline or purely online RL. Finally, our results illustrate the distinction between verifiable learning, the typical setting considered in online RL, and unverifiable learning, the setting often considered in offline RL, and show that there is a formal separation between these regimes.
|
https://proceedings.mlr.press/v202/wagner23a.html
|
https://proceedings.mlr.press/v202/wagner23a/wagner23a.pdf
|
https://openreview.net/forum?id=gLH40bhHpm
|
Fast Private Kernel Density Estimation via Locality Sensitive Quantization
|
https://proceedings.mlr.press/v202/wagner23a.html
|
Tal Wagner, Yonatan Naamad, Nina Mishra
|
https://proceedings.mlr.press/v202/wagner23a.html
|
ICML 2023
|
We study efficient mechanisms for differentially private kernel density estimation (DP-KDE). Prior work for the Gaussian kernel described algorithms that run in time exponential in the number of dimensions $d$. This paper breaks the exponential barrier, and shows how the KDE can privately be approximated in time linear in $d$, making it feasible for high-dimensional data. We also present improved bounds for low-dimensional data. Our results are obtained through a general framework, which we term Locality Sensitive Quantization (LSQ), for constructing private KDE mechanisms where existing KDE approximation techniques can be applied. It lets us leverage several efficient non-private KDE methods—like Random Fourier Features, the Fast Gauss Transform, and Locality Sensitive Hashing—and “privatize” them in a black-box manner. Our experiments demonstrate that our resulting DP-KDE mechanisms are fast and accurate on large datasets in both high and low dimensions.
|
https://proceedings.mlr.press/v202/walker23a.html
|
https://proceedings.mlr.press/v202/walker23a/walker23a.pdf
|
https://openreview.net/forum?id=m21SgZnBWZ
|
Investigating the Role of Model-Based Learning in Exploration and Transfer
|
https://proceedings.mlr.press/v202/walker23a.html
|
Jacob C Walker, Eszter Vértes, Yazhe Li, Gabriel Dulac-Arnold, Ankesh Anand, Theophane Weber, Jessica B Hamrick
|
https://proceedings.mlr.press/v202/walker23a.html
|
ICML 2023
|
State of the art reinforcement learning has enabled training agents on tasks of ever increasing complexity. However, the current paradigm tends to favor training agents from scratch on every new task or on collections of tasks with a view towards generalizing to novel task configurations. The former suffers from poor data efficiency while the latter is difficult when test tasks are out-of-distribution. Agents that can effectively transfer their knowledge about the world pose a potential solution to these issues. In this paper, we investigate transfer learning in the context of model-based agents. Specifically, we aim to understand where exactly environment models have an advantage and why. We find that a model-based approach outperforms controlled model-free baselines for transfer learning. Through ablations, we show that both the policy and dynamics model learnt through exploration matter for successful transfer. We demonstrate our results across three domains which vary in their requirements for transfer: in-distribution procedural (Crafter), in-distribution identical (RoboDesk), and out-of-distribution (Meta-World). Our results show that intrinsic exploration combined with environment models present a viable direction towards agents that are self-supervised and able to generalize to novel reward functions.
|
https://proceedings.mlr.press/v202/wan23a.html
|
https://proceedings.mlr.press/v202/wan23a/wan23a.pdf
|
https://openreview.net/forum?id=25fe54GXLo
|
UPSCALE: Unconstrained Channel Pruning
|
https://proceedings.mlr.press/v202/wan23a.html
|
Alvin Wan, Hanxiang Hao, Kaushik Patnaik, Yueyang Xu, Omer Hadad, David Güera, Zhile Ren, Qi Shan
|
https://proceedings.mlr.press/v202/wan23a.html
|
ICML 2023
|
As neural networks grow in size and complexity, inference speeds decline. To combat this, one of the most effective compression techniques – channel pruning – removes channels from weights. However, for multi-branch segments of a model, channel removal can introduce inference-time memory copies. In turn, these copies increase inference latency – so much so that the pruned model can be slower than the unpruned model. As a workaround, pruners conventionally constrain certain channels to be pruned together. This fully eliminates memory copies but, as we show, significantly impairs accuracy. We now have a dilemma: Remove constraints but increase latency, or add constraints and impair accuracy. In response, our insight is to reorder channels at export time, (1) reducing latency by reducing memory copies and (2) improving accuracy by removing constraints. Using this insight, we design a generic algorithm UPSCALE to prune models with any pruning pattern. By removing constraints from existing pruners, we improve ImageNet accuracy for post-training pruned models by 2.1 points on average – benefiting DenseNet (+16.9), EfficientNetV2 (+7.9), and ResNet (+6.2). Furthermore, by reordering channels, UPSCALE improves inference speeds by up to 2x over a baseline export.
|
https://proceedings.mlr.press/v202/wan23b.html
|
https://proceedings.mlr.press/v202/wan23b/wan23b.pdf
|
https://openreview.net/forum?id=JxtE52KIBR
|
Poisoning Language Models During Instruction Tuning
|
https://proceedings.mlr.press/v202/wan23b.html
|
Alexander Wan, Eric Wallace, Sheng Shen, Dan Klein
|
https://proceedings.mlr.press/v202/wan23b.html
|
ICML 2023
|
Instruction-tuned LMs such as ChatGPT, FLAN, and InstructGPT are finetuned on datasets that contain user-submitted examples, e.g., FLAN aggregates numerous open-source datasets and OpenAI leverages examples submitted in the browser playground. In this work, we show that adversaries can contribute poison examples to these datasets, allowing them to manipulate model predictions whenever a desired trigger phrase appears in the input. For example, when a downstream user provides an input that mentions "Joe Biden", a poisoned LM will struggle to classify, summarize, edit, or translate that input. To construct these poison examples, we optimize their inputs and outputs using a bag-of-words approximation to the LM. We evaluate our method on open-source instruction-tuned LMs. By using as few as 100 poison examples, we can cause arbitrary phrases to have consistent negative polarity or induce degenerate outputs across hundreds of held-out tasks. Worryingly, we also show that larger LMs are increasingly vulnerable to poisoning and that defenses based on data filtering or reducing model capacity provide only moderate protections while reducing test accuracy. Notice: This paper contains tasks with obscene content.
|
https://proceedings.mlr.press/v202/wan23c.html
|
https://proceedings.mlr.press/v202/wan23c/wan23c.pdf
|
https://openreview.net/forum?id=yXt417LWyI
|
SeMAIL: Eliminating Distractors in Visual Imitation via Separated Models
|
https://proceedings.mlr.press/v202/wan23c.html
|
Shenghua Wan, Yucen Wang, Minghao Shao, Ruying Chen, De-Chuan Zhan
|
https://proceedings.mlr.press/v202/wan23c.html
|
ICML 2023
|
Model-based imitation learning (MBIL) is a popular reinforcement learning method that improves sample efficiency on high-dimension input sources, such as images and videos. Following the convention of MBIL research, existing algorithms are highly deceptive by task-irrelevant information, especially moving distractors in videos. To tackle this problem, we propose a new algorithm - named Separated Model-based Adversarial Imitation Learning (SeMAIL) - decoupling the environment dynamics into two parts by task-relevant dependency, which is determined by agent actions, and training separately. In this way, the agent can imagine its trajectories and imitate the expert behavior efficiently in task-relevant state space. Our method achieves near-expert performance on various visual control tasks with complex observations and the more challenging tasks with different backgrounds from expert observations.
|
https://proceedings.mlr.press/v202/wan23d.html
|
https://proceedings.mlr.press/v202/wan23d/wan23d.pdf
|
https://openreview.net/forum?id=M6n9VdvThQ
|
Multiplier Bootstrap-based Exploration
|
https://proceedings.mlr.press/v202/wan23d.html
|
Runzhe Wan, Haoyu Wei, Branislav Kveton, Rui Song
|
https://proceedings.mlr.press/v202/wan23d.html
|
ICML 2023
|
Despite the great interest in the bandit problem, designing efficient algorithms for complex models remains challenging, as there is typically no analytical way to quantify uncertainty. In this paper, we propose Multiplier Bootstrap-based Exploration (MBE), a novel exploration strategy that is applicable to any reward model amenable to weighted loss minimization. We prove both instance-dependent and instance-independent rate-optimal regret bounds for MBE in sub-Gaussian multi-armed bandits. With extensive simulation and real-data experiments, we show the generality and adaptivity of MBE.
|
https://proceedings.mlr.press/v202/wan23e.html
|
https://proceedings.mlr.press/v202/wan23e/wan23e.pdf
|
https://openreview.net/forum?id=Vqj48SbjJt
|
Bandit Multi-linear DR-Submodular Maximization and Its Applications on Adversarial Submodular Bandits
|
https://proceedings.mlr.press/v202/wan23e.html
|
Zongqi Wan, Jialin Zhang, Wei Chen, Xiaoming Sun, Zhijie Zhang
|
https://proceedings.mlr.press/v202/wan23e.html
|
ICML 2023
|
We investigate the online bandit learning of the monotone multi-linear DR-submodular functions, designing the algorithm $\mathtt{BanditMLSM}$ that attains $O(T^{2/3}\log T)$ of $(1-1/e)$-regret. Then we reduce submodular bandit with partition matroid constraint and bandit sequential monotone maximization to the online bandit learning of the monotone multi-linear DR-submodular functions, attaining $O(T^{2/3}\log T)$ of $(1-1/e)$-regret in both problems, which improve the existing results. To the best of our knowledge, we are the first to give a sublinear regret algorithm for the submodular bandit with partition matroid constraint. A special case of this problem is studied by Streeter et al.(2009). They prove a $O(T^{4/5})$ $(1-1/e)$-regret upper bound. For the bandit sequential submodular maximization, the existing work proves an $O(T^{2/3})$ regret with a suboptimal $1/2$ approximation ratio (Niazadeh et al. 2021).
|
https://proceedings.mlr.press/v202/wang23a.html
|
https://proceedings.mlr.press/v202/wang23a/wang23a.pdf
|
https://openreview.net/forum?id=tbeLou0v9M
|
Tight Regret Bounds for Single-pass Streaming Multi-armed Bandits
|
https://proceedings.mlr.press/v202/wang23a.html
|
Chen Wang
|
https://proceedings.mlr.press/v202/wang23a.html
|
ICML 2023
|
Regret minimization in streaming multi-armed bandits (MABs) has been studied extensively, and recent work has shown that algorithms with $o(K)$ memory have to incur $\Omega(T^{2/3})$ regret, where $K$ and $T$ are the numbers of arms and trials. However, the previous best regret upper bound is still $O(K^{1/3} T^{2/3}\log^{1/3}(T))$, which is achieved by the simple uniform exploration algorithm. In this paper, we close this gap and complete the picture of regret minimization in single-pass streaming MABs. We first improve the regret lower bound to $\Omega(K^{1/3}T^{2/3})$ for algorithms with $o(K)$ memory. We then show that the $\log^{1/3}(T)$ factor is not necessary by designing algorithms with at most $O(\log^*(K))$-arm memory and achieve $O(K^{1/3}T^{2/3})$ expected regret based on streaming $\varepsilon$-best arm algorithms. We further tested the empirical performances of our algorithms on simulated MABs instances, where the proposed algorithms outperform the benchmark uniform exploration algorithm by a large margin and, on occasion, reduce the regret by up to 70%.
|
https://proceedings.mlr.press/v202/wang23b.html
|
https://proceedings.mlr.press/v202/wang23b/wang23b.pdf
|
https://openreview.net/forum?id=RHjLINycna
|
Improved Active Multi-Task Representation Learning via Lasso
|
https://proceedings.mlr.press/v202/wang23b.html
|
Yiping Wang, Yifang Chen, Kevin Jamieson, Simon Shaolei Du
|
https://proceedings.mlr.press/v202/wang23b.html
|
ICML 2023
|
To leverage the copious amount of data from source tasks and overcome the scarcity of the target task samples, representation learning based on multi-task pretraining has become a standard approach in many applications. However, up until now, most existing works design a source task selection strategy from a purely empirical perspective. Recently, Chen et al., 2022 gave the first active multi-task representation learning (A-MTRL) algorithm which adaptively samples from source tasks and can provably reduce the total sample complexity using the L2-regularized-target-source-relevance parameter $\nu^2$. But their work is theoretically suboptimal in terms of total source sample complexity and is less practical in some real-world scenarios where sparse training source task selection is desired. In this paper, we address both issues. Specifically, we show the strict dominance of the L1-regularized-relevance-based ($\nu^1$-based) strategy by giving a lower bound for the $\nu^2$-based strategy. When $\nu^1$ is unknown, we propose a practical algorithm that uses the LASSO program to estimate $\nu^1$. Our algorithm successfully recovers the optimal result in the known case. In addition to our sample complexity results, we also characterize the potential of our $\nu^1$-based strategy in sample-cost-sensitive settings. Finally, we provide experiments on real-world computer vision datasets to illustrate the effectiveness of our proposed method.
|
https://proceedings.mlr.press/v202/wang23c.html
|
https://proceedings.mlr.press/v202/wang23c/wang23c.pdf
|
https://openreview.net/forum?id=YxpkGn5Oly
|
Tilted Sparse Additive Models
|
https://proceedings.mlr.press/v202/wang23c.html
|
Yingjie Wang, Hong Chen, Weifeng Liu, Fengxiang He, Tieliang Gong, Youcheng Fu, Dacheng Tao
|
https://proceedings.mlr.press/v202/wang23c.html
|
ICML 2023
|
Additive models have been burgeoning in data analysis due to their flexible representation and desirable interpretability. However, most existing approaches are constructed under empirical risk minimization (ERM), and thus perform poorly in situations where average performance is not a suitable criterion for the problems of interest, e.g., data with complex non-Gaussian noise, imbalanced labels or both of them. In this paper, a novel class of sparse additive models is proposed under tilted empirical risk minimization (TERM), which addresses the deficiencies in ERM by imposing tilted impact on individual losses, and is flexibly capable of achieving a variety of learning objectives, e.g., variable selection, robust estimation, imbalanced classification and multiobjective learning. On the theoretical side, a learning theory analysis which is centered around the generalization bound and function approximation error bound (under some specific data distributions) is conducted rigorously. On the practical side, an accelerated optimization algorithm is designed by integrating Prox-SVRG and random Fourier acceleration technique. The empirical assessments verify the competitive performance of our approach on both synthetic and real data.
|
https://proceedings.mlr.press/v202/wang23d.html
|
https://proceedings.mlr.press/v202/wang23d/wang23d.pdf
|
https://openreview.net/forum?id=a032h8Jb9I
|
From Hypergraph Energy Functions to Hypergraph Neural Networks
|
https://proceedings.mlr.press/v202/wang23d.html
|
Yuxin Wang, Quan Gan, Xipeng Qiu, Xuanjing Huang, David Wipf
|
https://proceedings.mlr.press/v202/wang23d.html
|
ICML 2023
|
Hypergraphs are a powerful abstraction for representing higher-order interactions between entities of interest. To exploit these relationships in making downstream predictions, a variety of hypergraph neural network architectures have recently been proposed, in large part building upon precursors from the more traditional graph neural network (GNN) literature. Somewhat differently, in this paper we begin by presenting an expressive family of parameterized, hypergraph-regularized energy functions. We then demonstrate how minimizers of these energies effectively serve as node embeddings that, when paired with a parameterized classifier, can be trained end-to-end via a supervised bilevel optimization process. Later, we draw parallels between the implicit architecture of the predictive models emerging from the proposed bilevel hypergraph optimization, and existing GNN architectures in common use. Empirically, we demonstrate state-of-the-art results on various hypergraph node classification benchmarks. Code is available at https://github.com/yxzwang/PhenomNN.
|
https://proceedings.mlr.press/v202/wang23e.html
|
https://proceedings.mlr.press/v202/wang23e/wang23e.pdf
|
https://openreview.net/forum?id=UwEMFweZmC
|
A Closer Look at Self-Supervised Lightweight Vision Transformers
|
https://proceedings.mlr.press/v202/wang23e.html
|
Shaoru Wang, Jin Gao, Zeming Li, Xiaoqin Zhang, Weiming Hu
|
https://proceedings.mlr.press/v202/wang23e.html
|
ICML 2023
|
Self-supervised learning on large-scale Vision Transformers (ViTs) as pre-training methods has achieved promising downstream performance. Yet, how much these pre-training paradigms promote lightweight ViTs’ performance is considerably less studied. In this work, we develop and benchmark several self-supervised pre-training methods on image classification tasks and some downstream dense prediction tasks. We surprisingly find that if proper pre-training is adopted, even vanilla lightweight ViTs show comparable performance to previous SOTA networks with delicate architecture design. It breaks the recently popular conception that vanilla ViTs are not suitable for vision tasks in lightweight regimes. We also point out some defects of such pre-training, e.g., failing to benefit from large-scale pre-training data and showing inferior performance on data-insufficient downstream tasks. Furthermore, we analyze and clearly show the effect of such pre-training by analyzing the properties of the layer representation and attention maps for related models. Finally, based on the above analyses, a distillation strategy during pre-training is developed, which leads to further downstream performance improvement for MAE-based pre-training. Code is available at https://github.com/wangsr126/mae-lite.
|
https://proceedings.mlr.press/v202/wang23f.html
|
https://proceedings.mlr.press/v202/wang23f/wang23f.pdf
|
https://openreview.net/forum?id=gIJGcGoz44
|
PreNAS: Preferred One-Shot Learning Towards Efficient Neural Architecture Search
|
https://proceedings.mlr.press/v202/wang23f.html
|
Haibin Wang, Ce Ge, Hesen Chen, Xiuyu Sun
|
https://proceedings.mlr.press/v202/wang23f.html
|
ICML 2023
|
The wide application of pre-trained models is driving the trend of once-for-all training in one-shot neural architecture search (NAS). However, training within a huge sample space damages the performance of individual subnets and requires much computation to search for a optimal model. In this paper, we present PreNAS, a search-free NAS approach that accentuates target models in one-shot training. Specifically, the sample space is dramatically reduced in advance by a zero-cost selector, and weight-sharing one-shot training is performed on the preferred architectures to alleviate update conflicts. Extensive experiments have demonstrated that PreNAS consistently outperforms state-of-the-art one-shot NAS competitors for both Vision Transformer and convolutional architectures, and importantly, enables instant specialization with zero search cost. Our code is available at https://github.com/tinyvision/PreNAS.
|
https://proceedings.mlr.press/v202/wang23g.html
|
https://proceedings.mlr.press/v202/wang23g/wang23g.pdf
|
https://openreview.net/forum?id=LX3VAhXNTw
|
Adversarial Policies Beat Superhuman Go AIs
|
https://proceedings.mlr.press/v202/wang23g.html
|
Tony Tong Wang, Adam Gleave, Tom Tseng, Kellin Pelrine, Nora Belrose, Joseph Miller, Michael D Dennis, Yawen Duan, Viktor Pogrebniak, Sergey Levine, Stuart Russell
|
https://proceedings.mlr.press/v202/wang23g.html
|
ICML 2023
|
We attack the state-of-the-art Go-playing AI system KataGo by training adversarial policies against it, achieving a $>$97% win rate against KataGo running at superhuman settings. Our adversaries do not win by playing Go well. Instead, they trick KataGo into making serious blunders. Our attack transfers zero-shot to other superhuman Go-playing AIs, and is comprehensible to the extent that human experts can implement it without algorithmic assistance to consistently beat superhuman AIs. The core vulnerability uncovered by our attack persists even in KataGo agents adversarially trained to defend against our attack. Our results demonstrate that even superhuman AI systems may harbor surprising failure modes. Example games are available https://goattack.far.ai/.
|
https://proceedings.mlr.press/v202/wang23h.html
|
https://proceedings.mlr.press/v202/wang23h/wang23h.pdf
|
https://openreview.net/forum?id=PsQJm6lG3s
|
On Regularization and Inference with Label Constraints
|
https://proceedings.mlr.press/v202/wang23h.html
|
Kaifu Wang, Hangfeng He, Tin D. Nguyen, Piyush Kumar, Dan Roth
|
https://proceedings.mlr.press/v202/wang23h.html
|
ICML 2023
|
Prior knowledge and symbolic rules in machine learning are often expressed in the form of label constraints, especially in structured prediction problems. In this work, we compare two common strategies for encoding label constraints in a machine learning pipeline, regularization with constraints and constrained inference, by quantifying their impact on model performance. For regularization, we show that it narrows the generalization gap by precluding models that are inconsistent with the constraints. However, its preference for small violations introduces a bias toward a suboptimal model. For constrained inference, we show that it reduces the population risk by correcting a model’s violation, and hence turns the violation into an advantage. Given these differences, we further explore the use of two approaches together and propose conditions for constrained inference to compensate for the bias introduced by regularization, aiming to improve both the model complexity and optimal risk.
|
https://proceedings.mlr.press/v202/wang23i.html
|
https://proceedings.mlr.press/v202/wang23i/wang23i.pdf
|
https://openreview.net/forum?id=vYYjcdVqtG
|
Policy Gradient in Robust MDPs with Global Convergence Guarantee
|
https://proceedings.mlr.press/v202/wang23i.html
|
Qiuhao Wang, Chin Pang Ho, Marek Petrik
|
https://proceedings.mlr.press/v202/wang23i.html
|
ICML 2023
|
Robust Markov decision processes (RMDPs) provide a promising framework for computing reliable policies in the face of model errors. Many successful reinforcement learning algorithms build on variations of policy-gradient methods, but adapting these methods to RMDPs has been challenging. As a result, the applicability of RMDPs to large, practical domains remains limited. This paper proposes a new Double-Loop Robust Policy Gradient (DRPG), the first generic policy gradient method for RMDPs. In contrast with prior robust policy gradient algorithms, DRPG monotonically reduces approximation errors to guarantee convergence to a globally optimal policy in tabular RMDPs. We introduce a novel parametric transition kernel and solve the inner loop robust policy via a gradient-based method. Finally, our numerical results demonstrate the utility of our new algorithm and confirm its global convergence properties.
|
https://proceedings.mlr.press/v202/wang23j.html
|
https://proceedings.mlr.press/v202/wang23j/wang23j.pdf
|
https://openreview.net/forum?id=GdkwSGTpbC
|
Adaptive Smoothing Gradient Learning for Spiking Neural Networks
|
https://proceedings.mlr.press/v202/wang23j.html
|
Ziming Wang, Runhao Jiang, Shuang Lian, Rui Yan, Huajin Tang
|
https://proceedings.mlr.press/v202/wang23j.html
|
ICML 2023
|
Spiking neural networks (SNNs) with biologically inspired spatio-temporal dynamics demonstrate superior energy efficiency on neuromorphic architectures. Error backpropagation in SNNs is prohibited by the all-or-none nature of spikes. The existing solution circumvents this problem by a relaxation on the gradient calculation using a continuous function with a constant relaxation de- gree, so-called surrogate gradient learning. Nevertheless, such a solution introduces additional smoothing error on spike firing which leads to the gradients being estimated inaccurately. Thus, how to adaptively adjust the relaxation degree and eliminate smoothing error progressively is crucial. Here, we propose a methodology such that training a prototype neural network will evolve into training an SNN gradually by fusing the learnable relaxation degree into the network with random spike noise. In this way, the network learns adaptively the accurate gradients of loss landscape in SNNs. The theoretical analysis further shows optimization on such a noisy network could be evolved into optimization on the embedded SNN with shared weights progressively. Moreover, The experiments on static images, dynamic event streams, speech, and instrumental sounds show the proposed method achieves state-of-the-art performance across all the datasets with remarkable robustness on different relaxation degrees.
|
https://proceedings.mlr.press/v202/wang23k.html
|
https://proceedings.mlr.press/v202/wang23k/wang23k.pdf
|
https://openreview.net/forum?id=Fl9q5z40e3
|
CircuitNet: A Generic Neural Network to Realize Universal Circuit Motif Modeling
|
https://proceedings.mlr.press/v202/wang23k.html
|
Yansen Wang, Xinyang Jiang, Kan Ren, Caihua Shan, Xufang Luo, Dongqi Han, Kaitao Song, Yifei Shen, Dongsheng Li
|
https://proceedings.mlr.press/v202/wang23k.html
|
ICML 2023
|
The successes of artificial neural networks (ANNs) are largely attributed to mimicking the human brain structures. Recent advances in neuroscience revealed that neurons interact with each other through various kinds of connectivity patterns to process information, in which the common connectivity patterns are also called circuit motifs. However, many existing ANNs can only model one or two circuit motifs in their architectures, so that their performance may drastically vary among different types of machine learning tasks. In this paper, we propose a new type of neural network inspired by the architectures of neuronal circuits, namely Circuit Neural Network (CircuitNet). In CircuitNet, a group of densely connected neurons, namely circuit motif unit (CMU), form the basic unit of the network, which is capable of modeling universal circuit motifs by adjusting the weights within the CMUs. Compared with traditional feed-forward networks, CircuitNet has the ability to model more types of neuron connections such as feed-back and lateral motifs. Inspired by the locally dense and globally sparse structure of the human brain, several iterations of signal transmission among different CMUs are achieved by sparse connections through the input ports and output ports of different CMUs. Experiments have demonstrated that CircuitNet can outperform popular neural network architectures in function approximation, reinforcement learning, image classification, and time series forecasting tasks.
|
https://proceedings.mlr.press/v202/wang23l.html
|
https://proceedings.mlr.press/v202/wang23l/wang23l.pdf
|
https://openreview.net/forum?id=PeaecS9hm2
|
Generalized Polyak Step Size for First Order Optimization with Momentum
|
https://proceedings.mlr.press/v202/wang23l.html
|
Xiaoyu Wang, Mikael Johansson, Tong Zhang
|
https://proceedings.mlr.press/v202/wang23l.html
|
ICML 2023
|
In machine learning applications, it is well known that carefully designed learning rate (step size) schedules can significantly improve the convergence of commonly used first-order optimization algorithms. Therefore how to set step size adaptively becomes an important research question. A popular and effective method is the Polyak step size, which sets step size adaptively for gradient descent or stochastic gradient descent without the need to estimate the smoothness parameter of the objective function. However, there has not been a principled way to generalize the Polyak step size for algorithms with momentum accelerations. This paper presents a general framework to set the learning rate adaptively for first-order optimization methods with momentum, motivated by the derivation of Polyak step size. It is shown that the resulting techniques are much less sensitive to the choice of momentum parameter and may avoid the oscillation of the heavy-ball method on ill-conditioned problems. These adaptive step sizes are further extended to the stochastic settings, which are attractive choices for stochastic gradient descent with momentum. Our methods are demonstrated to be more effective for stochastic gradient methods than prior adaptive step size algorithms in large-scale machine learning tasks.
|
https://proceedings.mlr.press/v202/wang23m.html
|
https://proceedings.mlr.press/v202/wang23m/wang23m.pdf
|
https://openreview.net/forum?id=o6YrAc8XRm
|
Near-Minimax-Optimal Risk-Sensitive Reinforcement Learning with CVaR
|
https://proceedings.mlr.press/v202/wang23m.html
|
Kaiwen Wang, Nathan Kallus, Wen Sun
|
https://proceedings.mlr.press/v202/wang23m.html
|
ICML 2023
|
In this paper, we study risk-sensitive Reinforcement Learning (RL), focusing on the objective of Conditional Value at Risk (CVaR) with risk tolerance $\tau$. Starting with multi-arm bandits (MABs), we show the minimax CVaR regret rate is $\Omega(\sqrt{\tau^{-1}AK})$, where $A$ is the number of actions and $K$ is the number of episodes, and that it is achieved by an Upper Confidence Bound algorithm with a novel Bernstein bonus. For online RL in tabular Markov Decision Processes (MDPs), we show a minimax regret lower bound of $\Omega(\sqrt{\tau^{-1}SAK})$ (with normalized cumulative rewards), where $S$ is the number of states, and we propose a novel bonus-driven Value Iteration procedure. We show that our algorithm achieves the optimal regret of $\widetilde O(\sqrt{\tau^{-1}SAK})$ under a continuity assumption and in general attains a near-optimal regret of $\widetilde O(\tau^{-1}\sqrt{SAK})$, which is minimax-optimal for constant $\tau$. This improves on the best available bounds. By discretizing rewards appropriately, our algorithms are computationally efficient.
|
https://proceedings.mlr.press/v202/wang23n.html
|
https://proceedings.mlr.press/v202/wang23n/wang23n.pdf
|
https://openreview.net/forum?id=891ytYlYgB
|
FedHPO-Bench: A Benchmark Suite for Federated Hyperparameter Optimization
|
https://proceedings.mlr.press/v202/wang23n.html
|
Zhen Wang, Weirui Kuang, Ce Zhang, Bolin Ding, Yaliang Li
|
https://proceedings.mlr.press/v202/wang23n.html
|
ICML 2023
|
Research in the field of hyperparameter optimization (HPO) has been greatly accelerated by existing HPO benchmarks. Nonetheless, existing efforts in benchmarking all focus on HPO for traditional learning paradigms while ignoring federated learning (FL), a promising paradigm for collaboratively learning models from dispersed data. In this paper, we first identify some uniqueness of federated hyperparameter optimization (FedHPO) from various aspects, showing that existing HPO benchmarks no longer satisfy the need to study FedHPO methods. To facilitate the research of FedHPO, we propose and implement a benchmark suite FedHPO-Bench that incorporates comprehensive FedHPO problems, enables flexible customization of the function evaluations, and eases continuing extensions. We conduct extensive experiments based on FedHPO-Bench to provide the community with more insights into FedHPO. We open-sourced FedHPO-Bench at https://github.com/alibaba/FederatedScope/tree/master/benchmark/FedHPOBench.
|
https://proceedings.mlr.press/v202/wang23o.html
|
https://proceedings.mlr.press/v202/wang23o/wang23o.pdf
|
https://openreview.net/forum?id=9wP94EI8mk
|
A/B Testing in Network Data with Covariate-Adaptive Randomization
|
https://proceedings.mlr.press/v202/wang23o.html
|
Jialu Wang, Ping Li, Feifang Hu
|
https://proceedings.mlr.press/v202/wang23o.html
|
ICML 2023
|
Users linked together through a network often tend to have similar behaviors. This phenomenon is usually known as network interaction. Users’ characteristics, the covariates, are often correlated with their outcomes. Therefore, one should incorporate both the covariates and the network information in a carefully designed randomization to improve the estimation of the average treatment effect (ATE) in network A/B testing. In this paper, we propose a new adaptive procedure to balance both the network and the covariates. We show that the imbalance measures with respect to the covariates and the network are $O_p(1)$. We also demonstrate the relationships between the improved balances and the increased efficiency in terms of the mean square error (MSE). Numerical studies demonstrate the advanced performance of the proposed procedure regarding the greater comparability of the treatment groups and the reduction of MSE for estimating the ATE.
|
https://proceedings.mlr.press/v202/wang23p.html
|
https://proceedings.mlr.press/v202/wang23p/wang23p.pdf
|
https://openreview.net/forum?id=4IzEmHLono
|
Learning Belief Representations for Partially Observable Deep RL
|
https://proceedings.mlr.press/v202/wang23p.html
|
Andrew Wang, Andrew C Li, Toryn Q. Klassen, Rodrigo Toro Icarte, Sheila A. Mcilraith
|
https://proceedings.mlr.press/v202/wang23p.html
|
ICML 2023
|
Many important real-world Reinforcement Learning (RL) problems involve partial observability and require policies with memory. Unfortunately, standard deep RL algorithms for partially observable settings typically condition on the full history of interactions and are notoriously difficult to train. We propose a novel deep, partially observable RL algorithm based on modelling belief states — a technique typically used when solving tabular POMDPs, but that has traditionally been difficult to apply to more complex environments. Our approach simplifies policy learning by leveraging state information at training time, that may not be available at deployment time. We do so in two ways: first, we decouple belief state modelling (via unsupervised learning) from policy optimization (via RL); and second, we propose a representation learning approach to capture a compact set of reward-relevant features of the state. Experiments demonstrate the efficacy of our approach on partially observable domains requiring information seeking and long-term memory.
|
https://proceedings.mlr.press/v202/wang23q.html
|
https://proceedings.mlr.press/v202/wang23q/wang23q.pdf
|
https://openreview.net/forum?id=p5ZMcFXKvm
|
Warm-Start Actor-Critic: From Approximation Error to Sub-optimality Gap
|
https://proceedings.mlr.press/v202/wang23q.html
|
Hang Wang, Sen Lin, Junshan Zhang
|
https://proceedings.mlr.press/v202/wang23q.html
|
ICML 2023
|
Warm-Start reinforcement learning (RL), aided by a prior policy obtained from offline training, is emerging as a promising RL approach for practical applications. Recent empirical studies have demonstrated that the performance of Warm-Start RL can be improved quickly in some cases but become stagnant in other cases, especially when the function approximation is used. To this end, the primary objective of this work is to build a fundamental understanding on ”whether and when online learning can be significantly accelerated by a warm-start policy from offline RL?”. Specifically, we consider the widely used Actor-Critic (A-C) method with a prior policy. We first quantify the approximation errors in the Actor update and the Critic update, respectively. Next, we cast the Warm-Start A-C algorithm as Newton’s method with perturbation, and study the impact of the approximation errors on the finite-time learning performance with inaccurate Actor/Critic updates. Under some general technical conditions, we derive the upper bounds, which shed light on achieving the desired finite-learning performance in the Warm-Start A-C algorithm. In particular, our findings reveal that it is essential to reduce the algorithm bias in online learning. We also obtain lower bounds on the sub-optimality gap of the Warm-Start A-C algorithm to quantify the impact of the bias and error propagation.
|
https://proceedings.mlr.press/v202/wang23r.html
|
https://proceedings.mlr.press/v202/wang23r/wang23r.pdf
|
https://openreview.net/forum?id=ABzDOXlxf0
|
Slot-VAE: Object-Centric Scene Generation with Slot Attention
|
https://proceedings.mlr.press/v202/wang23r.html
|
Yanbo Wang, Letao Liu, Justin Dauwels
|
https://proceedings.mlr.press/v202/wang23r.html
|
ICML 2023
|
Slot attention has shown remarkable object-centric representation learning performance in computer vision tasks without requiring any supervision. Despite its object-centric binding ability brought by compositional modelling, as a deterministic module, slot attention lacks the ability to generate novel scenes. In this paper, we propose the Slot-VAE, a generative model that integrates slot attention with the hierarchical VAE framework for object-centric structured scene generation. For each image, the model simultaneously infers a global scene representation to capture high-level scene structure and object-centric slot representations to embed individual object components. During generation, slot representations are generated from the global scene representation to ensure coherent scene structures. Our extensive evaluation of the scene generation ability indicates that Slot-VAE outperforms slot representation-based generative baselines in terms of sample quality and scene structure accuracy.
|
https://proceedings.mlr.press/v202/wang23s.html
|
https://proceedings.mlr.press/v202/wang23s/wang23s.pdf
|
https://openreview.net/forum?id=Bxfp0zWygq
|
DIVISION: Memory Efficient Training via Dual Activation Precision
|
https://proceedings.mlr.press/v202/wang23s.html
|
Guanchu Wang, Zirui Liu, Zhimeng Jiang, Ninghao Liu, Na Zou, Xia Hu
|
https://proceedings.mlr.press/v202/wang23s.html
|
ICML 2023
|
Activation compressed training provides a solution towards reducing the memory cost of training deep neural networks (DNNs). However, state-of-the-art work combines a search of quantization bit-width with the training, which makes the procedure complicated and less transparent. To this end, we propose a simple and effective method to compress DNN training. Our method is motivated by an instructive observation: DNN backward propagation mainly utilizes the low-frequency component (LFC) of the activation maps, while the majority of memory is for caching the high-frequency component (HFC) during the training. This indicates the HFC of activation maps is highly redundant and compressible, which inspires our proposed Dual Activation Precision (DIVISION). During the training, DIVISION preserves a high-precision copy of LFC and compresses the HFC into a light-weight copy with low numerical precision. This can significantly reduce the memory cost while maintaining a competitive model accuracy. Experiment results show DIVISION has better comprehensive performance than state-of-the-art methods, including over 10x compression of activation maps and competitive training throughput, without loss of model accuracy. The source code is available at https://github.com/guanchuwang/division.
|
https://proceedings.mlr.press/v202/wang23t.html
|
https://proceedings.mlr.press/v202/wang23t/wang23t.pdf
|
https://openreview.net/forum?id=w2Vrl0zlzA
|
CocktailSGD: Fine-tuning Foundation Models over 500Mbps Networks
|
https://proceedings.mlr.press/v202/wang23t.html
|
Jue Wang, Yucheng Lu, Binhang Yuan, Beidi Chen, Percy Liang, Christopher De Sa, Christopher Re, Ce Zhang
|
https://proceedings.mlr.press/v202/wang23t.html
|
ICML 2023
|
Distributed training of foundation models, especially large language models (LLMs), is communication-intensive and so has heavily relied on centralized data centers with fast interconnects. Can we train on slow networks and unlock the potential of decentralized infrastructure for foundation models? In this paper, we propose CocktailSGD, a novel communication-efficient training framework that combines three distinct compression techniques – random sparsification, top-K sparsification, and quantization – to achieve much greater compression than each individual technique alone. We justify the benefit of such a hybrid approach through a theoretical analysis of convergence. Empirically, we show that CocktailSGD achieves up to 117$\times$ compression in fine-tuning LLMs up to 20 billion parameters without hurting convergence. On a 500Mbps network, CocktailSGD only incurs $\sim$1.2$\times$ slowdown compared with data center networks.
|
https://proceedings.mlr.press/v202/wang23u.html
|
https://proceedings.mlr.press/v202/wang23u/wang23u.pdf
|
https://openreview.net/forum?id=oeAhgeKFEw
|
Magneto: A Foundation Transformer
|
https://proceedings.mlr.press/v202/wang23u.html
|
Hongyu Wang, Shuming Ma, Shaohan Huang, Li Dong, Wenhui Wang, Zhiliang Peng, Yu Wu, Payal Bajaj, Saksham Singhal, Alon Benhaim, Barun Patra, Zhun Liu, Vishrav Chaudhary, Xia Song, Furu Wei
|
https://proceedings.mlr.press/v202/wang23u.html
|
ICML 2023
|
A big convergence of model architectures across language, vision, speech, and multimodal is emerging. However, under the same name ”Transformers”, the above areas use different implementations for better performance, e.g., Post-LayerNorm for BERT, and Pre-LayerNorm for GPT and vision Transformers. We call for the development of Foundation Transformer for true general-purpose modeling, which serves as a go-to architecture for various tasks and modalities with guaranteed training stability. In this work, we introduce a Transformer variant, named Magneto, to fulfill the goal. Specifically, we propose Sub-LayerNorm for good expressivity, and the initialization strategy theoretically derived from DeepNet for stable scaling up. Extensive experiments demonstrate its superior performance and better stability than the de facto Transformer variants designed for various applications, including language modeling (i.e., BERT, and GPT), machine translation, vision pretraining (i.e., BEiT), speech recognition, and multimodal pretraining (i.e., BEiT-3).
|
https://proceedings.mlr.press/v202/wang23v.html
|
https://proceedings.mlr.press/v202/wang23v/wang23v.pdf
|
https://openreview.net/forum?id=1lqOZrdXeG
|
Direct Parameterization of Lipschitz-Bounded Deep Networks
|
https://proceedings.mlr.press/v202/wang23v.html
|
Ruigang Wang, Ian Manchester
|
https://proceedings.mlr.press/v202/wang23v.html
|
ICML 2023
|
This paper introduces a new parameterization of deep neural networks (both fully-connected and convolutional) with guaranteed $\ell^2$ Lipschitz bounds, i.e. limited sensitivity to input perturbations. The Lipschitz guarantees are equivalent to the tightest-known bounds based on certification via a semidefinite program (SDP). We provide a “direct” parameterization, i.e., a smooth mapping from $\mathbb R^N$ onto the set of weights satisfying the SDP-based bound. Moreover, our parameterization is complete, i.e. a neural network satisfies the SDP bound if and only if it can be represented via our parameterization. This enables training using standard gradient methods, without any inner approximation or computationally intensive tasks (e.g. projections or barrier terms) for the SDP constraint. The new parameterization can equivalently be thought of as either a new layer type (the sandwich layer), or a novel parameterization of standard feedforward networks with parameter sharing between neighbouring layers. A comprehensive set of experiments on image classification shows that sandwich layers outperform previous approaches on both empirical and certified robust accuracy. Code is available at https://github.com/acfr/LBDN.
|
https://proceedings.mlr.press/v202/wang23w.html
|
https://proceedings.mlr.press/v202/wang23w/wang23w.pdf
|
https://openreview.net/forum?id=y6gg68aGiq
|
Tighter Information-Theoretic Generalization Bounds from Supersamples
|
https://proceedings.mlr.press/v202/wang23w.html
|
Ziqiao Wang, Yongyi Mao
|
https://proceedings.mlr.press/v202/wang23w.html
|
ICML 2023
|
In this work, we present a variety of novel information-theoretic generalization bounds for learning algorithms, from the supersample setting of Steinke & Zakynthinou (2020)—the setting of the “conditional mutual information” framework. Our development exploits projecting the loss pair (obtained from a training instance and a testing instance) down to a single number and correlating loss values with a Rademacher sequence (and its shifted variants). The presented bounds include square-root bounds, fast-rate bounds, including those based on variance and sharpness, and bounds for interpolating algorithms etc. We show theoretically or empirically that these bounds are tighter than all information-theoretic bounds known to date on the same supersample setting.
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.