Temporal cues improve long-term memory retrieval in AI agents
RL-Guided Temporal Localization for Dual-Channel Retrieval in Long-Horizon Agent Memory
Artificial Intelligence
Summary
Finding the right memories in a huge collection that grows over time is hard, especially when many memories sound alike but happened at different times. The authors created a way for AI agents to first guess when the important memories happened and then use that timing to find the right information. This approach helps avoid mixing up similar memories from the wrong time and makes the AI better at answering questions about past events. Their method works better than earlier ones on tests involving daily-like long memory.
What this means in practice
- •For conversational ai developers: Enhance AI agents that recall long-term conversations by precisely narrowing down when related information was stored for better response accuracy.
- •For personal assistant developers: Improve digital assistants’ ability to retrieve user-specific events from a growing timeline by integrating time-aware retrieval to reduce irrelevant memory matches.$Commercial implications: Enables more contextually accurate personal assistants that can better recall and recommend based on user history, a marketable feature for consumer devices and services.
Authors
Dabin Sheng, Zhe Wu, Jinming Zhao, Junliang Xing, Yuanchun Shi
Abstract
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 temporal constraints at the retrieval stage, where candidate selection occurs - the root cause of the displacement. Inspired by the temporal contiguity effect in human episodic memory, we propose TIDER (Temporal Interval-driven Dual-channel Evidence Retrieval), a time-aware agentic retrieval framework. Central to TIDER is a reinforcement learning (RL)-trained temporal interval locator that infers the temporal windows in which the evidence lies. RL optimizes window localization as a discrete, non-differentiable retrieval decision, using composite rewards to convert sparse answer supervision into fine-grained retrieval feedback. These temporal windows first define the temporal retrieval channel; the retrieved candidates are then merged with results from the semantic channel and reranked by proximity to the windows. This temporal grounding enables TIDER to retrieve temporally aligned evidence that would otherwise be drowned out by the volume of semantically similar yet temporally mismatched memories. With a lightweight 1.7B locator, TIDER achieves state-of-the-art accuracy on both long-horizon memory benchmarks - 61.52% on LifeBench (+8.82%) and 78.96% on LoCoMo (+3.66%) - which indicates that temporal cues are crucial for pinpointing target evidence amid a vast pool of semantically similar historical events in long-term daily-life memory.