AI Paper Atlas

Research map / Large language models / Retrieval and knowledge

BrowseNet: Knowledge Graph-Based Associative Memory for Contextual Information Retrieval

PAVAN KUMAR S, Kiran Kumar Nakka, C Vamshi Krishna Reddy, Divyateja Pasupuleti, Prakhar Agarwal, Harpinder Jot Singh, Anshu Avinash, Nirav Pravinbhai Bhatt

ICLR 2026 · Large language models › Retrieval and knowledge · approach: RAG robustness and evaluation

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
query-specific graph-of-chunks traversal instead of dense retrieval alone (BrowseNet vs NV-Embed-v2)retrieval · 2WikiMQARecall@5: 76.72 -> 93.30Improved
BrowseNet instead of HippoRAG-2retrieval · MuSiQueRecall@5: 74.20 -> 73.91No clear effect
BrowseNet instead of NV-Embed-v2retrieval · HotpotQARecall@2: 83.95 -> 83.85No clear effect
BrowseNet instead of HippoRAG-2answer generation · MuSiQueExact Match: 35.00 -> 41.60Improved
BrowseNet instead of HippoRAG-2answer generation · 2WikiMQAExact Match: 60.50 -> 63.90Improved
use graph-of-chunks to restrict corpus (ablation)BrowseNet · MuSiQue (100 questions)Recall@5 and time: 70.75 -> 75.08; 177 s -> 115 sImproved
lower ColBERT synonymity threshold to 0.7 (ablation)BrowseNet · Average retrievalRecall@5: 87.87 -> 87.72No clear effect
increase n_subgraphs to 10 or 15 (ablation)BrowseNet · Average retrievalRecall@5: 87.87 -> 87.71No clear effect
use Claude-3.7-Sonnet as query decomposer instead of GPT-4o (ablation)BrowseNet · Average retrievalRecall@5: 87.87 -> 86.37Hurt
replace NV-Embed-v2 encoder with GTE-Qwen2 (7B) (ablation)BrowseNet · Average retrievalRecall@5: 87.87 -> 81.90Hurt
use BrowseNet instead of NaiveRAG with Granite-125M encoderretrieval · Average Recall@5Recall@5: 71.76 -> 81.39Improved
answer with gpt-3.5-turbo instead of gpt-4o-mini (ablation)BrowseNet · Average QAExact Match: 55.90 -> 47.97Hurt

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