E-mem: Multi-agent based Episodic Context Reconstruction for LLM Agent Memory
arXiv:2601.21714v2 Announce Type: replace
Abstract: The evolution of Large Language Model (LLM) agents towards System~2 reasoning, characterized by deliberative, high-precision problem-solving, requires maintaining rigorous logical integrity over exte…