Qianxi Li

MSc student at the University of Alberta

I obtained my bachelor of science honors in computing science degree in June 2022, from the University of Alberta. Then started my master of science in computing science thesis-based program at the University of Alberta in September 2022. I am currently working at the Intelligent Robot Learning Lab (IRLL), under the supervision of Dr. Matthew E. Taylor. I am also doing my internship at Huawei Noah's Ark Lab as an Associate Reinforcement Learning Researcher from May 2023.

My current research directions include:
1. Applied ML/RL. Solving real-world problems using machine learning (ML) and reinforcement learning (RL) approaches.
2. Human-in-the-Loop (HITL) ML/RL. An area which incorporates humans as one essential component of the learning process. I am especially interested in the intersection area of Large Language Models, Human-in-the-Loop methods and Deep RL. I believe that to make AI a better tool for us, human advice and feedback must be taken by the AI system, with HITL, we can make an AI system safer, train faster and better align with humans' needs and preferences.
3. Foundation Models (LLMs). Multi-Modality Pretrained Models. Cross Domain/Modalities LLM Finetuning. Foundation Models for Planning and Control Problems.
4. Multi-task RL. Sample efficiency is important in RL! And the features a model learned can be partially used for different tasks. Also as humans, as we grow up, we are more and more unlikely to learn something from the absolute beginning, right? We (partially) utilize the knowledge from a learned skill!

Want to know more about my experience and what projects I'm working on? Visit the experience and projects sections on the top navigation bar.