RL-Guided Temporal Localization for Dual-Channel Retrieval in Long-Horizon Agent Memory
Time-aware retrieval is crucial for long-horizon LLM agents to effectively leverage ever-growing long-term memory. Existing methods focus on lexical or semantic similarity, leading to target memories being crowded out by temporally mismatched yet semantically similar candidates. Even time-aware approaches inject temporal cues only during memory organization or answer generation, overlooking…
We haven't written up this one. arXiv cs.AI has the full story — the link below goes straight to it.
