AI Paper Atlas

Research map / Large language models / Agents, code and math

SP-Mind: An Autonomous Reasoning Agent for Spatial Proteomics Analysis

YuCheng Yuan, Yuanfeng Ji, Zhongxiao Li, Ruijiang Li

ICML 2026 · Large language models › Agents, code and math · approach: LLM-driven embodied and driving agents · 0 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
add expert-curated skill templates (ablation)SP-Mind · SP-Bench averageexecution accuracy (%): 62.4 -> 68.9Improved
add expert-curated skill templates (ablation)SP-Mind · SP-Bench Advanced tierexecution accuracy (%): 44.4 -> 61.9Improved
add expert-curated skill templates (ablation)SP-Mind · SP-Bench Basic tierexecution accuracy (%): 95.0 -> 95.8No clear effect
add expert-curated skill templates (ablation)SP-Mind · CRC-CODEX QuantificationPearson r: 0.945 -> 0.953No clear effect
add expert-curated skill templates (ablation)SP-Mind · Cell Annotation PDAC IMCCyteOnto GHK similarity: 0.762 -> 0.648Hurt
add expert-curated skill templates (ablation)SP-Mind · Cell Annotation (4 datasets, average)CyteOnto GHK similarity: 0.657 -> 0.681Improved
add expert-curated skill templates (ablation)SP-Mind · Cell Annotation cHL 2 MIBI 5CyteOnto GHK similarity: 0.672 -> 0.765Improved
equip generalist agent with spatial proteomics toolkitBiomni · SP-Bench averageexecution accuracy (%): 19.3 -> 55.4Improved
equip generalist agent with spatial proteomics toolkitToolUniverse · SP-Bench averageexecution accuracy (%): 15.0 -> 50.5Improved
use SP-Mind agent instead of Biomni-SP with same toolsBiomni-SP · SP-Bench Advanced tierexecution accuracy (%): 36.5 -> 61.9Improved
use SP-Mind instead of specialized ML baseline ASTIR on cell annotationASTIR · Cell Annotation (4 datasets, average)CyteOnto GHK similarity: 0.440 -> 0.681Improved
longer multi-stage pipelines (Challenging tier) for all agentsgeneralist agents · SP-Bench Challenging tierexecution accuracy (%): 0.0 for AutoGen, Biomni, ToolUniverse; 33.3 for SP-MindHurt

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