Enhong Chen
5 papers · Latest:
Learning How and What to Memorize: Cognition-Inspired Two-Stage Optimization for Evolving Memory
MemCoE is a cognition-inspired two-stage optimization framework that learns how to organize and what to update in LLM memory for better personalization.
Understanding DNNs in Feature Interaction Models: A Dimensional Collapse Perspective
This paper shows DNNs in feature interaction models mitigate dimensional collapse, improving representation robustness and clarifying their role.
StepPO: Step-Aligned Policy Optimization for Agentic Reinforcement Learning
StepPO introduces step-aligned policy optimization for Agentic RL, shifting from token-level to step-level MDPs to enhance LLM agent capabilities.
Tango: Taming Visual Signals for Efficient Video Large Language Models
Tango optimizes token pruning in Video LLMs by improving attention selection and similarity clustering, achieving significant speedup with minimal performance loss.
A Survey on Multimodal Large Language Models
This paper surveys recent advances in Multimodal Large Language Models (MLLMs), highlighting their architectures, training, capabilities, and future research directions.
📬 Weekly AI Paper Digest
Get the top 10 AI/ML arXiv papers from the week — summarized, scored, and delivered to your inbox every Monday.