LExGraph
A graph-based memory framework that incrementally organizes dialogue into interconnected memory nodes. Evidence is retrieved through hybrid retrieval combined with neighbor expansion, supporting long-horizon memory in exploration tasks.
Timeline (last 90 days)
Research from the MBZUAI NLP team proposes the graph-guided benchmark LExMem and the graph-based memory framework LExGraph for memory modeling and evaluation in long-horizon exploration tasks.
LExGraph incrementally organizes dialogues into interconnected memory nodes and obtains evidence through hybrid retrieval and neighbor expansion.
In LExMem benchmark experiments, LExGraph achieves better performance than existing retrieval-based and memory-based systems and generalizes effectively to external benchmarks.
In LExGraph, hybrid retrieval is responsible for locating seeds among memory nodes, and neighbor expansion supplements this with structural associations, forming a two-stage evidence acquisition process of "locate first, then diffuse".
All Facts (4)
Research from the MBZUAI NLP team proposes the graph-guided benchmark LExMem and the graph-based memory framework LExGraph for memory modeling and evaluation in long-horizon exploration tasks.
50%UnverifiedLExGraph incrementally organizes dialogues into interconnected memory nodes and obtains evidence through hybrid retrieval and neighbor expansion.
50%UnverifiedIn LExMem benchmark experiments, LExGraph achieves better performance than existing retrieval-based and memory-based systems and generalizes effectively to external benchmarks.
50%UnverifiedIn LExGraph, hybrid retrieval is responsible for locating seeds among memory nodes, and neighbor expansion supplements this with structural associations, forming a two-stage evidence acquisition process of "locate first, then diffuse".
50%