AI Paper Atlas

Research map / Large language models / Retrieval and knowledge

Hybrid Deep Searcher: Scalable Parallel and Sequential Search Reasoning

Dayoon Ko, Jihyuk Kim, Haeju Park, Sohyeon Kim, Dahyun Lee, Yongrae Jo, Gunhee Kim, Moontae Lee, Kyungjae Lee

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
train on HDS-QA hybrid parallel+sequential trajectories vs single-query trajectories (ablation)HybridDeepSearcher · MuSiQueAcc: 21.6 -> 35.1Improved
train on hybrid trajectories vs single-query trajectories (ablation)HybridDeepSearcher · FanOutQAAcc: 8.7 -> 20.0Improved
train on hybrid trajectories vs single-query trajectories (ablation)HybridDeepSearcher · BrowseComp subsetAcc: 6.4 -> 13.3Improved
apply GRPO on top of the supervised model (ablation)HybridDeepSearcher · FRAMESAcc: 54.0 -> 57.2Improved
apply GRPO on top of the supervised model (ablation)HybridDeepSearcher · five benchmarksAUC (efficiency): 0.44 -> 0.42 (FRAMES)Hurt
apply GRPO on top of the supervised model (ablation)HybridDeepSearcher · BrowseComp subset# search turns: 6.0 -> 6.7Hurt
supervised fine-tuning on HDS-QA instead of RL-trained baselines (Qwen2.5-7B-Instruct backbone)Qwen2.5-7B-Instruct · FanOutQAF1: 28.2 -> 37.4Improved
add HDS-QA hybrid fine-tuning to Qwen3-8B vs prompt-based Search-o1Qwen3-8B · FanOutQAF1: 26.7 -> 44.1Improved
hybrid search vs best parallel baseline RAG-R1HybridDeepSearcher · FanOutQAevidence coverage rate: 53.2 -> 61.0Improved
increase search turn limit from 1 to 8HybridDeepSearcher · BrowseComp subsetF1: 4.0 -> 15.1Improved
increase search call limit from 2 to 16HybridDeepSearcher · BrowseComp subsetF1: 3.4 -> 15.1Improved
single-query fine-tuned variant vs prompt-based Search-o1 (ablation)HybridDeepSearcher single-query · FRAMESAcc: 48.6 -> 33.8Hurt

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