1 paper
Sihang Jiang, Lipeng Ma, Zhonghua Hong +9
Current LLM-based agents demonstrate strong performance in episodic task execution but remain constrained by static toolsets and episodic amnesia, failing to accumulate experience…