CV
Kaiwen Xiong
Summary
Fourth-year undergraduate at Shanghai Jiao Tong University. Currently working on AI.
Education
- B.S. in Automation2027Shanghai Jiao Tong University
Publications
- ClawArena-Team: Benchmarking Subagent Orchestration and Dynamic Workflows in Language-Model Agents2026arXiv preprintClawArena-Team measures how well a language-model agent manages specialized subagents through delegation, least-privilege access control, modality routing, and dynamic workflows.
- ClawArena: Benchmarking AI Agents in Evolving Information Environments2026arXiv preprintClawArena benchmarks whether persistent AI agents can maintain correct beliefs as information evolves across noisy, partial, and contradictory multi-channel evidence.
- MetaClaw: Just Talk -- An Agent That Meta-Learns and Evolves in the Wild2026arXiv preprintMetaClaw is an agent that meta-learns and evolves in the wild. It turns every live conversation into a learning signal by placing the model behind a proxy that injects relevant skills at each turn and meta-learns from accumulated experience, enabling continuous improvement through real-world deployment without requiring a GPU cluster.
- Alignment Tipping Process: How Self-Evolution Pushes LLM Agents Off the Rails2026arXiv preprintWe identify the Alignment Tipping Process, a critical post-deployment risk unique to self-evolving LLM agents. We formalize and analyze Alignment Tipping Process through two complementary paradigms: Self-Interested Exploration, where repeated high-reward deviations induce individual behavioral drift, and Imitative Strategy Diffusion, where deviant behaviors spread across multi-agent systems.
- Agent0-VL: Exploring Self-Evolving Agent for Tool-Integrated Vision-Language Reasoning2026International Conference on Machine Learning (ICML), Oral
- An Adaptive Sliding Window Estimator for Positioning of Unmanned Aerial Vehicle Using a Single Anchor2025IEEE Sensors JournalWe propose an adaptive sliding window estimator equipped with an estimation reliability evaluator, where the states, noise covariance matrices, and aerial drag are estimated simultaneously.
Presentations
- Talk 1 on Relevant Topic in Your Field2012UC San Francisco, Department of TestingSan Francisco, CA, USA
- Tutorial 1 on Relevant Topic in Your Field2013UC-Berkeley Institute for Testing ScienceBerkeley, CA, USA
- Talk 2 on Relevant Topic in Your Field2014London School of TestingLondon, UK
- Conference Proceeding talk 3 on Relevant Topic in Your Field2014Testing Institute of America 2014 Annual ConferenceLos Angeles, CA, USA
Teaching
- Teaching experience 12014University 1, DepartmentRole: Undergraduate course
- Teaching experience 22015University 1, DepartmentRole: Workshop
Portfolio
- Portfolio item number 1
