Liang Qiu

My name is Liang Qiu. My research interests include Natural Language Processing and Conversational AI. My long-term research goal is to improve the EQ (aligning AI with human values, mental state modeling, and social relation inference) and IQ (reasoning with deep learning) of agents. I'm currently focused on using RL-based alignment algorithms with agentic and reasoning critics to address scalable oversight challenges. My approach involves developing a “mixture of experts” agentic critic system that can generate structured rubrics, search and execute relevant tools or reasoning steps, act on the task context, and dispatch sub-tasks as needed. This critic provides reliable feedback signals to guide the LLM policy during RL post-training, enhancing its helpfulness, reasoning capability, and alignment with human preferences.

​​

I served as a reviewer of NeurIPS, ACL, EMNLP, SIGDIAL, NLPCC, IEEE/ACM Transactions on Audio, Speech, and Language Processing, Journal of Artificial Intelligence Research, etc.