NLP & Large Language Models 4: LLM Agents & Multi-Agent Systems
Wednesday, November 25, 2026, – (Australia/Sydney)
Clarendon Room B · Conference
Paper 515 - Execution-Guided Shared-Seed Multi-Agent Refinement for LLM-Based Data Science Code Generation
Authors: Xin Wang, Xuebin Zhang and Chunlong Fan
Paper 1241 - LATS-R: A Training-Free Self-Reflective Language Agent Tree Search Framework for Financial Directional Forecasting
Authors: Kanghui Wang, Wei Wang and Enxin Kang
Paper 1368 - External Emotion-Based Pursuit-Evasion Strategies for LLM Agents
Authors: Zicheng Zhai and Yu Qiao
Paper 982 - TEMPO: Hierarchical Phase Control Reduces Token Cost in Long-Horizon LLM Tool Use
Authors: Hu Yanxi, Zhao Yanjie, Huang Yiwen, Xu Fanyi, Di Mingyu, Zhang Baoming and Wang Chongjun
Paper 1320 - VitaMem: Validity-Guided Task Evolving Memory for Long-Horizon LLM Agents
Authors: Guixuan Zhang, Zhuo Wang, Xiadan Chen, Jie Liu and Shuwu Zhang
Paper 1328 - Beyond Which Model Wins: Capability Profiles and Failure Fingerprints of LLM Agents in Multi-Agent Games
Authors: Zhendong Li, Feimo Li, Min Chen, Tenghai Qiu and Zhiqiang Pu