Grounded verification of chemical and materials reasoning: detection is the bottleneck

2026-07-19 · Can Polat, Mustafa Kurban, Erchin Serpedin, Hasan Kurban arxiv

Language models are moving into chemistry and materials discovery workflows, where a wrong molecular formula, space group, or formation energy can silently propagate into downstream decisions. These confabulations hide inside fluent reasoning traces and concentrate on rare, long-tail entities, where model confidence is least trustworthy. Retrieving reference data for every prompt would catch them, but at a heavy coverage and abstention cost. We show that deterministic, database-grounded verification catches and repairs these errors selectively, and that the binding constraint is detection rather than repair. Our tiered verifier extracts each checkable claim, tests it against authoritative databases and physical law, and retrieves a reference value only when a check fails. Across four models and over five hundred prompts with pinned conditions, gated correction cuts the error rate of committed formulas from 22% to 4% with 3.2 times fewer retrievals than blanket augmentation, and it outperforms a conversational retrieval oracle when every answer, corrected or not, is scored. When a flag fires, repair almost always succeeds; the benefit reaches the final answer only where the verifier's scope covers it and where long-tail error exists. Checkable claims, checked cheaply, are a practical lever for trustworthy machine reasoning in chemistry.

📄 PDF Abstract BibTeX arXiv:2607.17417

Code (0)

등록된 구현이 없습니다.

Hierarchical Multi-agent Large Language Model Reasoning for Autonomous Functional Materials Discovery

2025-12-15 · Samuel Rothfarb, Megan C. Davis, Ivana Matanovic, Baikun Li 외 arxiv

Artificial intelligence is reshaping scientific exploration, but most methods automate procedural tasks without engaging in scientific reasoning, limiting autonomy in discovery. We introduce Materials Agents for Simulati…

Active Learning

Expanding the extreme-k dielectric materials space through physics-validated generative reasoning

2026-04-22 · Hossain Hridoy, Tahiya Chowdhury, Md Shafayat Hossain arxiv

The most technologically consequential materials are often the rarest: they occupy narrow regions of chemical space, obey competing physical constraints, and appear only sparsely in existing databases. High-kappa dielect…

MOF-Sleuth: Tool-Grounded Reward Alignment for Explainable Fine-Grained MOF CIF Auditing

2026-07-22 · Yu Liu, Zhiwei Yang, Diandian Guo, Kun Peng 외 arxiv

Large metal-organic framework (MOF) databases support simulation, screening, and machine learning through crystallographic information files (CIFs). Subtle chemical and structural errors in these inputs can compromise do…

Reinforcement Learning

Enhancing Materials Discovery with Valence Constrained Design in Generative Modeling

2025-07-26 · Mouyang Cheng, Weiliang Luo, Hao Tang, Bowen Yu 외 arxiv

Diffusion-based deep generative models have emerged as powerful tools for inverse materials design. Yet, many existing approaches overlook essential chemical constraints such as oxidation state balance, which can lead to…

Reasoning-Enhanced Large Language Models for Molecular Property Prediction

2025-10-11 · Jiaxi Zhuang, Yaorui Shi, Jue Hou, Yunong He 외 arxiv

Molecular property prediction is crucial for drug discovery and materials science, yet existing approaches suffer from limited interpretability, poor cross-task generalization, and lack of chemical reasoning capabilities…

Molecular Property Prediction Reinforcement Learning Drug Discovery