Liang, Yingbin
96 publications
NeurIPS
2025
Absorb and Converge: Provable Convergence Guarantee for Absorbing Discrete Diffusion Models
ICLR
2025
Broadening Target Distributions for Accelerated Diffusion Models via a Novel Analysis Approach
NeurIPS
2025
Multi-Head Transformers Provably Learn Symbolic Multi-Step Reasoning via Gradient Descent
NeurIPSW
2024
Enhancing Generalization in Sparse Mixture of Experts Models: The Case for Increased Expert Activation in Compositional Tasks
NeurIPS
2024
In-Context Learning with Representations: Contextual Generalization of Trained Transformers
ICMLW
2024
In-Context Learning with Representations: Contextual Generalization of Trained Transformers
NeurIPSW
2024
Model Developmental Safety: A Safety-Centric Method and Applications in Vision-Language Models
NeurIPS
2024
Training Dynamics of Transformers to Recognize Word Co-Occurrence via Gradient Flow Analysis
ICML
2023
A Near-Optimal Algorithm for Safe Reinforcement Learning Under Instantaneous Hard Constraints
ICML
2023
Generalized-Smooth Nonconvex Optimization Is as Efficient as Smooth Nonconvex Optimization
ICLR
2023
Theoretical Characterization of the Generalization Performance of Overfitted Meta-Learning
AISTATS
2021
Sample Complexity Bounds for Two Timescale Value-Based Reinforcement Learning Algorithms
AISTATS
2021
When Will Generative Adversarial Imitation Learning Algorithms Attain Global Convergence
AAAI
2021
Non-Asymptotic Convergence of Adam-Type Reinforcement Learning Algorithms Under Markovian Sampling
IJCAI
2020
Proximal Gradient Algorithm with Momentum and Flexible Parameter Restart for Nonconvex Optimization
JMLR
2017
A Nonconvex Approach for Phase Retrieval: Reshaped Wirtinger Flow and Incremental Algorithms