Evolutionary System Prompt Learning for Reinforcement Learning in LLMs
Conference on Neural Information Processing Systems (NeurIPS), 2026
Combining RL and Evolutionary Algorithms for more effective LLM Self-Improvement.
Research
I am broadly interested in building general-purpose agents in the digital and physical worlds, with a focus on recursive self improvement.
I work on self-improvement mechanisms from various perspectives, from reward modeling to world modeling to RL to automated AI research.
Selected Publications
2026
2025
2024
Copilot4D: Learning Unsupervised World Models for Autonomous Driving via Discrete Diffusion
International Conference on Learning Representations (ICLR), 2024
A foundation world model for self-driving that explicitly reasons in both 3D space and time.
2023
Towards Unsupervised Object Detection from LiDAR Point Clouds
Conference on Computer Vision and Pattern Recognition (CVPR), 2023
Self-supervised, scalable object discovery in the wild.
2021
World Model as a Graph: Learning Latent Landmarks for Planning
International Conference on Machine Learning (ICML), 2021 (Long Talk)
Unsupervised long-horizon planning via graph-structured world models.