Research
A collection of research that I am not too ashamed of
To me, research and entrepreneurship are quite alike: you are responsible for your work and your team while creating something new. The allure of grounded exploration is my antidote to modern or post-modern nihilism, which is why I have pursued both equally for the past five years.
Although I am now paving my way on the startup side, I remain passionate about reading and conducting research, and I hope to continue doing so in the future.
Research Interests
I am interested in decision-making that bridges artificial intelligence and societal needs: Computational models give abstraction, societal systems give databases and constraints, and decition-making with responsibility stands between. Here are some topics that recently went to my mind and my hand:
- How does the outdated data and real-world constraints occur in societal systems impose difficulties on reinforcement learning? And how to address these difficulties provably? [1]
- How to extract contextual variations and the implied preference in crafting incentives from Chinese environmental governance documents ? [Master Thesis]
Published Work
Distributionally Robust Constrained Reinforcement Learning under Strong Duality
Reinforcement Learning Journal, vol. 4, 2024, pp. 1793–1821 / the 1st Reinforcement Learning Conference, 2024
An Invariant Information Geometric Method for High-Dimensional Online Optimization
Proceedings of the 6th Annual Learning for Dynamics & Control Conference(L4DC), 2024
