|
                          Peng "Richard" Xia (夏鹏)
I am a Ph.D. student at Department of Computer Science, The University of North Carolina at Chapel Hill, advised by Prof. Huaxiu Yao.
I am also a student researcher at Google, working on autonomous agent design and Gemini Deep Research.
Previously, I conducted research at Tongyi Lab, Alibaba and Microsoft Research.
I build long-horizon LLM & VLM agents through mid/post-training (imitation or reinforcement learning) and data scaling. My research focuses on equipping autonomous self-evolving systems with long-term memory and complex tool-use capabilities to tackle open-ended workflows that require 10-1000s of human hours.
My work was accepted as oral or spotlight presentations at venues like ICLR (2025, 2026), NeurIPS (2025), ICML (2026), ACL (2026), received the Best Paper Runner-Up Award at ICLR MemAgents Workshop, and Oustanding Paper Award at ICLR RSI Workshop.
Email: richard.peng.xia AT gmail DOT com; pxia AT cs DOT unc DOT edu
 / 
 / 
 / 
 / 
 / 
 / 
|
📸 Caught on camera by Tars in San Diego
|
Agentic Evolution (skills, memory, tool-usage etc)
MetaClaw: Just Talk -- An Agent That Meta-Learns and Evolves in the Wild
Peng Xia, Jianwen Chen, Xinyu Yang, Haoqin Tu, Jiaqi Liu, Kaiwen Xiong, Siwei Han, Shi Qiu, Haonian Ji, Yuyin Zhou, Zeyu Zheng, Cihang Xie, Huaxiu Yao
arXiv preprint, 2026.
[Paper]
[Code]
|
SkillRL: Evolving Agents via Recursive Skill-Augmented Reinforcement Learning
Peng Xia, Jianwen Chen, Hanyang Wang, Jiaqi Liu, Kaide Zeng, Yu Wang, Siwei Han, Yiyang Zhou, Xujiang Zhao, Haifeng Chen, Zeyu Zheng, Cihang Xie, Huaxiu Yao
ICLR 2026 Workshop on MemAgents (Oral) Best Paper Award Runner-Up.
[Paper]
[Code]
|
Agentic Mid/Post-Training (data scaling)
WebWatcher: Breaking New Frontiers of Vision-Language Deep Research Agent
Xinyu Geng*, Peng Xia*, Zhen Zhang*, Xinyu Wang, Qiuchen Wang, Ruixue Ding, Chenxi Wang, Jialong Wu, Yida Zhao, Kuan Li, Yong Jiang, Pengjun Xie, Fei Huang, Jingren Zhou
International Conference on Learning Representations (ICLR), 2026.
[Paper]
[Code]
|
Invited Talks
-
Jan. 2026: MiroMind AI, Agent0 & Agent0-VL: Unleashing Self-Evolving Agents via Tool-Integrated Reasoning.
-
Jul. 2025: TechBeat, MMed-RAG: Versatile Multimodal RAG System for Medical Vision Language Models.
-
Apr. 2025: ICLR Oral Session, Massive Multimodal Interleaved Comprehension Benchmark For Large Vision-Language Models.
-
Dec. 2024: Cohere For AI, Reliable Multimodal RAG for Factuality in Medical Vision Language Models. [Video] [Cohere Event] [X/Twitter] [Linkedin]
|
|
Press
-
"MetaClaw" was covered by 36Kr, The Decoder. "SkillRL" was covered by 36Kr. "AutoResearchClaw" was covered by MIT Technology Review.
-
"Tongyi Deep Research" was covered by IBM News, China Economic News, South China Morning Post, ForkLog, Apidog, Medium, Towards AI, MarkTechPost, VentureBeat, Geeky Gadgets, ASO World, AIBase.
-
"WebWatcher" was covered by So Essentially, BYCLOUD AI, Beehiiv, Medium Towards Dev.
-
"MMed-RAG" was covered by Banff International Research Station, MarkTechPost, Moonlight Press, neptune.ai.
|
|
Selected Honors & Awards
-
ACL Oral Presentation (Top 3.3%), 2026
-
ICLR 2026 MemAgents Workshop Best Paper Award Runner-Up, 2026
-
ICLR 2026 RSI Workshop Outstanding Paper Award, 2026
-
ICML Spotlight Presentation (Top 2.2%), 2026
-
ICLR Oral Presentation (Top 1.1%), 2026
-
ICDM UGHS 2025 Rising Star Award & Best Poster Award, 2025
-
NeurIPS Spotlight Presentation (Top 3.2%), 2025
-
ICLR Oral Presentation (Top 1.8%), 2025
-
Stars of Tomorrow Excellent Intern Award, Microsoft Research, 2025
-
KDD 2025 Health Day Distinguished Vision Award, 2025
-
ICLR Travel Award, 2025
-
Third Place, Shanghai-HK Interdisciplinary Shared Tasks (Task 1), 2022
-
Second Price, The 3rd Huawei DIGIX AI Algorithm Contest, 2021
|
|
Academic Services
-
Area Chair: ACL Rolling Review (ARR)
-
Conference Reviewer: NeurIPS, ICML, ICLR, CVPR, ICCV, ACL Rolling Review (ARR), ECCV, MICCAI, WACV, AAAI, KDD, COLM
-
Journal Reviewer: IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), International Journal of Computer Vision (IJCV), IEEE Transactions on Medical Imaging (TMI), Cell Patterns, Knowledge-Based Systems (KBS), Expert Systems with Applications (ESWA), Pattern Recognition (PR), ACM Computing Surveys
-
Student Volunteer: EMNLP (2024)
-
Workshop Co-Organizer: ICML 2025 Workshop on Reliable and Responsible Foundation Models
|
Teaching
-
Teaching Assistant, DATA 110, School of Data Science and Society, UNC-Chapel Hill, 2026 Spring.
|
© Peng Xia | Last updated:
| |