AI Paper Atlas

Research map / Large language models / Retrieval and knowledge

Beyond Fact Retrieval: Episodic Memory for RAG with Generative Semantic Workspaces

Shreyas Rajesh, Pavan Holur, Chenda Duan, David Chong, Vwani Roychowdhury

AAAI 2026 · Large language models › Retrieval and knowledge · approach: Retrieval-augmented reasoning · 6 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
replace chunk/graph retrieval with GSW structured episodic memoryGPT-4o QA · EpBench-200 overallF1: 0.771 -> 0.850Improved
use GSW instead of HippoRAG2 on multi-cue queriesGPT-4o QA · EpBench-200 6+ cuesrecall: 0.675 -> 0.822Improved
use GSW at 10x corpus scaleGPT-4o QA · EpBench-2000 overallF1: 0.675 -> 0.773Improved
use entity-specific GSW summaries rather than chunksGPT-4o QA · EpBench-200context tokens: 8771 -> 3587 (vs embedding RAG); 51% fewer than GraphRAGImproved
GSW vs GraphRAG on queries with one cueGPT-4o QA · EpBench-200 1 cueF1: 0.625 -> 0.744Improved
GSW vs Embedding RAG on 1-cue precisionGPT-4o QA · EpBench-200 1 cueprecision: 0.745 -> 0.755No clear effect
GSW vs HippoRAG2 6+ cue precisionGPT-4o QA · EpBench-200 6+ cuesprecision: 0.940 -> 0.890Hurt
structured RAG baselines (GraphRAG, LightRAG) vs embedding RAGGPT-4o QA · EpBench-2000F1: 0.675 -> 0.544 / 0.494Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy