Agentic generation of verifiable rules for deterministic, self-expanding reaction classification

Source

• Source name: arXiv
• Source kind: paper
• Authors / organization: Daniel Armstrong, Maarten Dobbelaere, Valentas Olikauskas, Helena Avila, Octavian Susanu, Jérôme Waser, Philippe Schwaller
• Venue: arXiv: cs.AI
• Published: 2026-07-01
• Collected: 2026-07-02
• URL: http://arxiv.org/abs/2607.01061v1
• PDF: https://arxiv.org/pdf/2607.01061v1
• DOI: UNKNOWN

Plain-Language Brief

Researchers have developed an automated pipeline using large language models (LLMs) that can classify reactions in chemistry and generate rules for these reactions, expanding a standard taxonomy from 68 to 14,073 classes without human curation. The system is able to classify 97.7% of unseen reactions accurately, outperforming a leading proprietary classifier.

What the Document Says

• The research presents an automated pipeline for reaction classification in chemistry using LLMs.
• The pipeline generates rules for each transformation under a verification loop that tests the rules against the corpus of 665,901 US patent reactions.
• The standard taxonomy is expanded from 68 to 14,073 classes without human curation.
• The system uses a lightweight fingerprint classifier and achieves a classification accuracy of 97.7% on unseen reactions.
• The new method outperforms a leading proprietary classifier by resolving chemistry more finely and extending to chemistry outside its training distribution.

Key Concepts Mentioned

• computer-assisted synthesis planning
• large language models
• reaction rules
• chemistry
• US patent reactions
• taxonomy
• fingerprint classifier

Claims Made by the Document

• The automated pipeline can accurately classify reactions and generate rules for these reactions.
• The system expands the taxonomy without human curation.
• The new method outperforms a leading proprietary classifier.

Evidence Used

• Computer-assisted synthesis planning breaks target molecules into accessible precursors using large libraries of reaction rules.
• The pipeline expands a standard taxonomy from 68 to 14,073 classes without human curation.
• With a lightweight fingerprint classifier, it classifies 97.7% of unseen reactions.

Methods / Approach

UNKNOWN

Limitations Stated by the Authors

UNKNOWN

Limitations Noted by the Agent

Agent interpretation — not from the source.

UNKNOWN

Open Questions

UNKNOWN

Notes for Future Q&A

May include interpretation. Labeled where it does.

UNKNOWN

Raw Abstract / Excerpt

Computer-assisted synthesis planning breaks target molecules into accessible precursors using large libraries of reaction rules that assign each transformation a deterministic, interpretable label. But chemistry is long-tailed, making manual encoding intractable, and existing tools rely on fixed rulesets that cannot adapt to new chemistries. Here we present a fully automated pipeline in which a multi-agent framework of large language models (LLMs) classifies reactions and writes the rules themselves across 665,901 US patent reactions, generating each rule under a verification loop that tests it against the corpus. It expands a standard taxonomy from 68 to 14,073 classes without human curation. With a lightweight fingerprint classifier, it classifies 97.7% of unseen reactions, matching a leading proprietary classifier while resolving chemistry more finely and extending on demand to chemistry outside its training distribution. The result is a living reactivity database and a general route to turning generative models into reliable, self-expanding symbolic systems.