Zihao Tang / 唐子豪

Zihao Tang

唐子豪 · IshiKura-a

Hi, I'm Zihao. I am currently at Microsoft, working on LLM agents, AI memory, and agentic reinforcement learning.

I received my Master's degree from Zhejiang University, where I was fortunate to be advised by A.P. Kun Kuang and Prof. Fei Wu and to work with AI4GC Lab.

Microsoft · Zhejiang University

ScholarLinkedInEmail

Research interests

Agents that learn from experience.

LLM Agents

I care about agents that remain reliable when tasks require long-horizon planning, tool use, search, and code interaction rather than one-shot response generation.

AI Memory

I care about memory that is organized, selective, and useful for future decisions, not just a larger context window or similarity-based recall.

Agentic RL

I care about agents that improve from interaction and feedback, turning repeated experience into reusable procedures rather than isolated task traces.

Selected publications

Representative papers from recent work.

2026ACL

Mnemis: Dual-Route Retrieval on Hierarchical Graphs for Long-Term LLM Memory

Authors: Zihao Tang, Xin Yu, Ziyu Xiao, Zengxuan Wen, Zelin Li, Jiaxi Zhou, Hualei Wang, Haohua Wang, Haizhen Huang, Weiwei Deng, Feng Sun, Qi Zhang

Long-term memory retrieval for LLM agents; reaches 93.9 on LoCoMo and 91.6 on LongMemEval-S with GPT-4.1-mini.

PaperCode
2026ACL

TL; DR: Too Long, Do Re-weighting for Efficient LLM Reasoning Compression

Authors: Zhong-Zhi Li, Xiao Liang, Zihao Tang, Lei Ji, Peijie Wang, Haotian Xu, Xing W, Haizhen Huang, Weiwei Deng, Yeyun Gong, Zhijiang Guo, Xiao Liu, Fei Yin, Cheng-Lin Liu

Reasoning compression through Thinking Length Data Re-weighting, reducing output tokens by nearly 40% while maintaining accuracy.

PaperCode
2025arXiv

Sigma: Differential Rescaling of Query, Key and Value for Efficient Language Models

Authors: Zhenghao Lin, Zihao Tang, Xiao Liu, Yeyun Gong, Yi Cheng, Qi Chen, Hang Li, Ying Xin, Ziyue Yang, Kailai Yang, Yu Yan, Xiao Liang, Shuai Lu, Yiming Huang, Zheheng Luo, Lei Qu, Xuan Feng, Yaoxiang Wang, Yuqing Xia, Feiyang Chen, Yuting Jiang, Yasen Hu, Hao Ni, Binyang Li, Guoshuai Zhao, Jui-Hao Chiang, Zhongxin Guo, Chen Lin, Kun Kuang, Wenjie Li, Yelong Shen, Jian Jiao, Peng Cheng, Mao Yang

Efficient system-domain language modeling with DiffQKV attention for faster long-context inference.

Paper
2024arXiv

ModelGPT: Unleashing LLM's Capabilities for Tailored Model Generation

Authors: Zihao Tang, Zheqi Lv, Shengyu Zhang, Fei Wu, Kun Kuang

LLM-assisted model generation from user data or task descriptions.

PaperCode
2024ICLR

AuG-KD: Anchor-Based Mixup Generation for Out-of-Domain Knowledge Distillation

Authors: Zihao Tang, Zheqi Lv, Shengyu Zhang, Yifan Zhou, Xinyu Duan, Fei Wu, Kun Kuang

Data-free knowledge distillation under domain shift with uncertainty-guided anchors and mixup generation.

PaperCode
View all publications