Learning Theory

Universal Rates for Interactive Learning NeurIPS 2022 Unlabelled Sample Compression Schemes for Intersection-Closed Classes and Extremal Classes NeurIPS 2022 Unobserved Local Structures Make Compositional Generalization Hard EMNLP 2022 Unpacking Reward Shaping: Understanding the Benefits of Reward Engineering on Sample Complexity NeurIPS 2022 Using time-series privileged information for provably efficient learning of prediction models AISTATS 2022 Vanishing Curvature in Randomly Initialized Deep ReLU Networks AISTATS 2022 VC dimension of partially quantized neural networks in the overparametrized regime ICLR 2022 Verification and search algorithms for causal DAGs NeurIPS 2022 Verified Path Following Using Neural Control Lyapunov Functions CoRL 2022 Vision Transformers provably learn spatial structure NeurIPS 2022 Weighted Training for Cross-Task Learning ICLR 2022 What Happens after SGD Reaches Zero Loss? --A Mathematical Framework ICLR 2022 What is a Good Metric to Study Generalization of Minimax Learners? NeurIPS 2022 When are Local Queries Useful for Robust Learning? NeurIPS 2022 When Can Transformers Ground and Compose: Insights from Compositional Generalization Benchmarks EMNLP 2022 When Combinatorial Thompson Sampling meets Approximation Regret NeurIPS 2022 When Hardness of Approximation Meets Hardness of Learning JMLR 2022 When Is Partially Observable Reinforcement Learning Not Scary? COLT 2022 When More Data Hurts: A Troubling Quirk in Developing Broad-Coverage Natural Language Understanding Systems EMNLP 2022 Why Fair Labels Can Yield Unfair Predictions: Graphical Conditions for Introduced Unfairness AAAI 2022 Why Robust Generalization in Deep Learning is Difficult: Perspective of Expressive Power NeurIPS 2022 Wide Neural Networks Forget Less Catastrophically ICML 2022 Width is Less Important than Depth in ReLU Neural Networks COLT 2022 LeadCache: Regret-Optimal Caching in Networks NeurIPS 2021 A Biased Graph Neural Network Sampler with Near-Optimal Regret NeurIPS 2021