AI Paper Atlas

Research map / Probabilistic and scientific modeling / Molecules and proteins

AutoMat: Physics-Guided Agentic Reasoning for Solving Ill-Posed Inverse Microscopy Problems

Yaotian Yang, Yiwen Tang, Yizhe Chen, Xiao Chen, Jiangjie Qiu, Hao Xiong, Haoyu Yin, Zhiyao Luo, Yifei Zhang, Sijia Tao, Wentao Li, Qinghua Zhang and 5 more

ICML 2026 · Probabilistic and scientific modeling › Molecules and proteins · approach: Machine learning for physical simulation

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
use the LLM failure-aware controller instead of a hand-coded heuristic script with same tools (ablation)AutoMat · STEM2Mat-Bench overallformation energy MAE (meV/atom): 385.42 -> 321.57Improved
use the LLM controller instead of heuristic script on hardest tier (ablation)AutoMat · STEM2Mat-Bench Tier 3formation energy MAE (meV/atom): 585.00 -> 333.49Improved
use the LLM controller instead of heuristic script (ablation)AutoMat · STEM2Mat-Bench Tier 3structure success rate S.S. (%): 62.3 -> 73.1Improved
use the LLM controller instead of heuristic script (ablation)AutoMat · STEM2Mat-Bench overallcomposition correctness C.C. (%): 76.5 -> 83.2Improved
use the LLM controller instead of heuristic script (ablation)AutoMat · STEM2Mat-Bench overallstructure success rate S.S. (%): 74.2 -> 83.2Improved
swap DeepSeek backbone for GPT-4o (ablation)AutoMat · STEM2Mat-Bench avg over tiersformation energy MAE (meV/atom): 321.57 -> 323.25No clear effect
swap DeepSeek backbone for GPT-4o (ablation)AutoMat · STEM2Mat-Bench Tier 2in-plane lattice RMSD (A): 0.11 -> 0.12No clear effect
use AutoMat instead of off-the-shelf MicroscopyGPTAutoMat · STEM2Mat-Bench avg over tiersformation energy MAE (meV/atom): 620.22 -> 321.57Improved
use AutoMat instead of MicroscopyGPTAutoMat · STEM2Mat-Bench Tier 2structure success rate S.S. (%): 25.9 -> 84.0Improved
use AutoMat instead of general VLMs such as GPT-4.1 VisionAutoMat · STEM2Mat-Bench avg over tiersformation energy MAE (meV/atom): 2657.3 -> 321.57Improved

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