Nanxu Gong

I obtained my Bachelor degree from Sichuan University in 2024, where I was supervised by Prof. Chenwei Tang. I was interned at Microsoft Research Asia in the summer of 2025, advised by Dr. Haotian Li and Dr. Jianxun Lian.

My research interests lie in Large Language Model, Multi-Agent System, and Social Intelligence.

News

[2026-01] One papers is accepted by ICLR2026.

[2025-12] One papers is accepted by TKDD and one paper is accepted by TII.

[2025-09] Two papers are accepted by AAAI2026. Congratulations to my co-authors!

[2025-09] One paper is accepted by NeurIPS2025.

[2025-05] One paper is accepted by npj Artificial Intelligence.

[2025-04] I will join Microsoft Research Asia as a research intern during this summer.

[2025-04] One paper is accepted by IJCAI2025.

[2024-12] One paper is accepted by AAAI2025.

[2024-12] One paper is accepted by TIST.

	Sculpting Features from Noise: Reward-Guided Hierarchical Diffusion for Task-Optimal Feature Transformation Nanxu Gong^, Zijun Li^, Sixun Dong, Haoyue Bai, Wangyang Ying, Xinyuan Wang, Yanjie Fu NeurIPS, 2025 project page / arXiv We propose a reward-guided semi-autoregressive diffusion model to generate feature transformation functions that optimize downstream task performance.
	Unsupervised Feature Transformation via In-context Generation, Generator-critic LLM Agents, and Duet-play Teaming Nanxu Gong, Xinyuan Wang, Wangyang Ying, Haoyue Bai, Sixun Dong, Haifeng Chen, Yanjie Fu IJCAI, 2025 project page / arXiv We propose a generator-critic LLM agent framework to automatically generate feature transformation in an unsupervised manner.
	Evolutionary Large Language Model for Automated Feature Transformation Nanxu Gong, Chandan K Reddy, Wangyang Ying, Haifeng Chen, Yanjie Fu AAAI, 2025 project page / arXiv We integrate the evolutionary algorithm with the large language model to automatically generate feature transformation functions.
	Neuro-Symbolic Embedding for Short and Effective Feature Selection via Autoregressive Generation Nanxu Gong^, Wangyan Ying^, Dongjie Wang, Yanjie Fu ACM TIST project page / arXiv We formulate the feature selection task from an autoregressive generative perspective.