Bin Li

Latest

• Tackling Heavy-Tailed Q-Value Bias in Offline-to-Online Reinforcement Learning with Laplace-Robust Modeling
• Why Attention Patterns Exist: A Unifying Temporal Perspective Analysis
• AttentionPredictor: Temporal Patterns Matter for KV Cache Compression
• Benchmarking End-To-End Performance of AI-Based Chip Placement Algorithms
• LogicTree: Improving Complex Reasoning of LLMs via Instantiated Multi-step Synthetic Logical Data
• Accurate and Scalable Graph Neural Networks via Message Invariance
• Differentiable Integer Linear Programming
• Learning Robust Representations with Long-Term Information for Generalization in Visual Reinforcement Learning
• MILP-StuDio: MILP Instance Generation via Block Structure Decomposition
• Towards Next-Generation Logic Synthesis: A Scalable Neural Circuit Generation Framework
• Uncertainty-based Offline Variational Bayesian Reinforcement Learning for Robustness under Diverse Data Corruptions
• Coarse-to-Fine Highlighting: Reducing Knowledge Hallucination in Large Language Models
• Deep symbolic optimization for combinatorial optimization: Accelerating node selection by discovering potential heuristics
• Accelerating Data Generation for Neural Operators via Krylov Subspace Recycling (Spotlight)
• Rethinking Branching on Exact Combinatorial OPtimization Solver: The First Deep Symbolic Discovery Framework
• Learning Task-relevant Representations for Generalization via Characteristic Functions of Reward Sequence Distributions
• Learning Robust Policy against Disturbance in Transition Dynamics via State-Conservative Policy Optimization
• Sample-Efficient Reinforcement Learning via Conservative Model-Based Actor-Critic
• Interpreting the Latent Space of GANs via Measuring Decoupling