Nan Tang @ LAMDA, NJU-AI

汤楠 Nan Tang M.Sc. Student, LAMDA Group School of Artificial Intelligence Nanjing University, Nanjing 210023, China Supervisor: Professor Yang Yu Email: tangn@lamda.nju.edu.cn Google Scholar: Google Scholar

Biography

Currently I am a second year Ph.d student of School of Artificial Intelligence in Nanjing University and a member of LAMDA Group, led by professor Zhi-Hua Zhou. I got my B.Sc. degree in Kuang Yaming Honors School in June 2022 from Nanjing University, majored in statistics. In the same year, I was admitted to pursue my master degree in Nanjing University without entrance examination.

My research interests include Reinforcement Learning, Embodied AI for Robotics and World Model. Currently, I am working on learning latent world models for visual robotic manipulation and how to apply it to train an VLA using RL.

Conference Paper

Nan Tang , Jingcheng Pang, Guanlin Li, Chao Qian, Yang Yu. ReLAM: Learning Anticipation Model for Rewarding Visual Robotic Manipulation, In: Proceedings of the 43rd International Conference on Machine Learning (ICML'26) , Seoul, Korea, 2026. [paper]

Jingcheng Pang*, Nan Tang* , Kaiyuan Li, Yang Yu. Learning View-invariant World Models for Visual Robotic Manipulation, In: Proceedings of the 13th International Conference on Learning Representations (ICLR'25) , Singapore, 2025. [paper] [code]

Other Publications

Pengsen Liu*, Maoseng Zeng*, Nan Tang , Kaiyuan Li, Jingcheng Pang, Yunan Liu, Yang Yu. VLGOR: Visual-Language Knowledge Guided Offline Reinforcement Learning for Generalizable Agents, Preprint. [paper]

Jing-Cheng Pang, Si-Hang Yang, Kaiyuan Li, Xiong-Hui Chen, Nan Tang, Yang Yu. Kalm: Knowledgeable agents by offline reinforcement learning from large language model rollouts, In: Advances in Neural Information Processing Systems 37 (NeurIPS'24). [paper]

Projects and Contests

1. Participated in a quant project for market-trend prediction and trading-policy training:

• Utilized historical k-line data to build a world model for future market-trend prediction;
• Trained an RL trading-policy with the predicted trend;
• Achieved profits in both simulated and live trading backtests.

2. Achieve 4th place in the in the RL-Quant Contest organized by Kafang.

Correspondence

Email: tangn@lamda.nju.edu.cn

Address: National Key Laboratory for Novel Software Technology, Nanjing University, Xianlin Campus Mailbox 603, 163 Xianlin Avenue, Qixia District, Nanjing 210023, China