Jun Chen

Dual-Process Atomic Skill Learning: Decoupling Semantic Reasoning and Real-Time Control

Jun Chen, Erdent Bao, Wenlong Dong, Jierui Liu, Qi Cai, Hao Wan, et al.

Language-conditioned Imitation Learning (IL) is essential for enabling robots to perform complex tasks following natural language instructions. However, generalizing to multi-step compositional tasks remains a significant challenge. While hierarchical approaches attempt to addres…Read more

NFTR: From Provable Mode-Averaging to Geodesic Subgoal Selection in Offline Goal-Conditioned RL

Erdemt Bao, Xing Lei, Jun Chen

Hierarchical Implicit Q-Learning (HIQL), an offline goal-conditioned RL method, selects subgoals by value-function advantages alone. This rule has two coupled failure modes. Optimistic bias treats lucky stochastic outcomes as skillful choices, and mode collapse reduces a multi-mo…Read more

Oyster-II: Reinforcement Learning for Constructive Safety Alignment in Large Language Models

Jiyang Guan, Yong Xie, Jun Chen, Jiexi Liu, Zipeng Ye, Defeng Li, et al.

Large language models (LLMs) have demonstrated remarkable capabilities across diverse applications, yet ensuring their simultaneous safety, helpfulness, and trustworthiness remains a persistent challenge. Conventional refusal-oriented alignment strategies mitigate harmful content…Read more

Customizing Tactile Sensors via Machine Learning‐Driven Inverse Design (Adv. Sci. 29/2026)

Baocheng Wang, Depeng Kong, Zhiao He, Jikai Liang, Yuyao Lu, Zikang Deng, et al.