7 papers
HyMem: Hierarchical Context Management for Long-Horizon Agents via Information Isolation
XinQi Wang, Jinwei Xiao, Sijia Cui +4
Large language model (LLM) agents often perform poorly on complex, long-horizon tasks because their context becomes increasingly cluttered over time. As interactions accumulate, de…
Empowering LLMs with Parameterized Skills for Adversarial Long-Horizon Planning
Sijia Cui, Shuai Xu, Aiyao He +2
Recent advancements in Large Language Models(LLMs) have led to the development of LLM-based AI agents. A key challenge is the creation of agents that can effectively ground themsel…
Self-Guided Function Calling in Large Language Models via Stepwise Experience Recall
Sijia Cui, Aiyao He, Shuai Xu +5
Function calling enables large language models (LLMs) to interact with external systems by leveraging tools and APIs. When faced with multi-step tool usage, LLMs still struggle wit…
Strategy-Augmented Planning for Large Language Models via Opponent Exploitation
Shuai Xu, Sijia Cui, Yanna Wang +2
Efficiently modeling and exploiting opponents is a long-standing challenge in adversarial domains. Large Language Models (LLMs) trained on extensive textual data have recently demo…
TUMS: Enhancing Tool-use Abilities of LLMs with Multi-structure Handlers
Aiyao He, Sijia Cui, Shuai Xu +2
Recently, large language models(LLMs) have played an increasingly important role in solving a wide range of NLP tasks, leveraging their capabilities of natural language understandi…
Efficient Preference-based Reinforcement Learning via Aligned Experience Estimation
Fengshuo Bai, Rui Zhao, Hongming Zhang +5
Preference-based reinforcement learning (PbRL) has shown impressive capabilities in training agents without reward engineering. However, a notable limitation of PbRL is its depende…
Byzantine-Resilient Federated Learning at Edge
Youming Tao, Sijia Cui, Wenlu Xu +4
Both Byzantine resilience and communication efficiency have attracted tremendous attention recently for their significance in edge federated learning. However, most existing algori…