
When Can We Learn GeneralSum Markov Games with a Large Number of Players SampleEfficiently?
Multiagent reinforcement learning has made substantial empirical progre...
read it

CrossLingual Language Model MetaPretraining
The success of pretrained crosslingual language models relies on two es...
read it

Understanding the UnderCoverage Bias in Uncertainty Estimation
Estimating the data uncertainty in regression tasks is often done by lea...
read it

Policy Finetuning: Bridging SampleEfficient Offline and Online Reinforcement Learning
Recent theoretical work studies sampleefficient reinforcement learning ...
read it

Multimodal Trajectory Prediction for Autonomous Driving with Semantic Map and Dynamic Graph Attention Network
Predicting future trajectories of surrounding obstacles is a crucial tas...
read it

Exact Gap between Generalization Error and Uniform Convergence in Random Feature Models
Recent work showed that there could be a large gap between the classical...
read it

SampleEfficient Learning of Stackelberg Equilibria in GeneralSum Games
Real world applications such as economics and policy making often involv...
read it

Localized Calibration: Metrics and Recalibration
Probabilistic classifiers output confidence scores along with their pred...
read it

Don't Just Blame Overparametrization for Overconfidence: Theoretical Analysis of Calibration in Binary Classification
Modern machine learning models with high accuracy are often miscalibrate...
read it

NearOptimal Offline Reinforcement Learning via Double Variance Reduction
We consider the problem of offline reinforcement learning (RL) – a well...
read it

How Important is the TrainValidation Split in MetaLearning?
Metalearning aims to perform fast adaptation on a new task through lear...
read it

A Sharp Analysis of Modelbased Reinforcement Learning with SelfPlay
Modelbased algorithms—algorithms that decouple learning of the model an...
read it

Near Optimal Provable Uniform Convergence in OffPolicy Evaluation for Reinforcement Learning
The OffPolicy Evaluation aims at estimating the performance of target p...
read it

Towards Understanding Hierarchical Learning: Benefits of Neural Representations
Deep neural networks can empirically perform efficient hierarchical lear...
read it

NearOptimal Reinforcement Learning with SelfPlay
This paper considers the problem of designing optimal algorithms for rei...
read it

Provable SelfPlay Algorithms for Competitive Reinforcement Learning
Selfplay, where the algorithm learns by playing against itself without ...
read it

Taylorized Training: Towards Better Approximation of Neural Network Training at Finite Width
We propose Taylorized training as an initiative towards better understan...
read it

DirectedWeighting Group Lasso for Eltwise Blocked CNN Pruning
Eltwise layer is a commonly used structure in the multibranch deep lear...
read it

Beyond Linearization: On Quadratic and HigherOrder Approximation of Wide Neural Networks
Recent theoretical work has established connections between overparamet...
read it

Provably Efficient QLearning with Low Switching Cost
We take initial steps in studying PACMDP algorithms with limited adapti...
read it

Proximal algorithms for constrained composite optimization, with applications to solving lowrank SDPs
We study a family of (potentially nonconvex) constrained optimization p...
read it

Subgradient Descent Learns Orthogonal Dictionaries
This paper concerns dictionary learning, i.e., sparse coding, a fundamen...
read it

ProxQuant: Quantized Neural Networks via Proximal Operators
To make deep neural networks feasible in resourceconstrained environmen...
read it

Approximability of Discriminators Implies Diversity in GANs
While Generative Adversarial Networks (GANs) have empirically produced i...
read it

CirCNN: Accelerating and Compressing Deep Neural Networks Using BlockCirculantWeight Matrices
Largescale deep neural networks (DNNs) are both compute and memory inte...
read it

The Landscape of Empirical Risk for Nonconvex Losses
Most highdimensional estimation and prediction methods propose to minim...
read it
Yu Bai
is this you? claim profile