Yuan Cao

UNDERSTANDING TRANSFORMERS FOR TIME SERIES FORECASTING: A CASE STUDY ON MOIRAI

Transformers Trained via Gradient Descent Can Provably Learn a Class of Teacher Models

Towards Simple and Provable Parameter-Free Adaptive Gradient Methods

Learning Convolutional Representations via Generalized Stein’s Method: A Training-Free Approach

Understanding the Generalization of Stochastic Gradient Adam in Learning Neural Networks

On the Robustness of Transformers against Context Hijacking for Linear Classification

Towards Understanding Transformers in Learning Random Walks

On the Feature Learning in Diffusion Models

Transformer Learns Optimal Variable Selection in Group-Sparse Classification

Towards Understanding Token Selection in Self-Attention: Successes and Pitfalls in Learning Random Walks

On the Robustness of Transformers against Context Hijacking for Linear Classification

Understanding the Benefits of SimCLR Pre-Training in Two-Layer Convolutional Neural Networks

Can Transformers Perform PCA ?

Towards Simple and Provable Parameter-Free Adaptive Gradient Methods

Transformers Learn Bayesian Networks Autoregressively In-Context