Sign InOpen Brain
arXivPaperNeeds Review

DASyR-LLM: Domain-Aware Symbolic Regression with LLMs for Kinetic Model Discovery

DASyR-LLM adds model critique and candidate generation to symbolic regression, cutting search iterations in simulated chemistry studies without improving final fit.

arXiv · Aug 5, 2026
Open Source Open MarkdownOpen JSON
Source Summary

DASyR-LLM alternates symbolic regression with an LLM that critiques candidate equations and proposes replacements using chemical knowledge. Across **4 in silico case studies**, it cut iterations to the ground-truth model by **41.7–79.3%** and directly proposed the correct structure in over half of guided runs.

Practical Implication

Builders of research agents should separate numerical search from domain critique, then feed the strongest candidates back into both. The ablations suggest a smaller LLM can retain much of the discovery efficiency, so model scale need not be the first lever.

Agent-Ready Context
DASyR-LLM alternates symbolic regression with an LLM that critiques candidate equations and proposes replacements using chemical knowledge. Across **4 in silico case studies**, it cut iterations to the ground-truth model by **41.7–79.3%** and directly proposed the correct structure in over half of guided runs.

Builders of research agents should separate numerical search from domain critique, then feed the strongest candidates back into both. The ablations suggest a smaller LLM can retain much of the discovery efficiency, so model scale need not be the first lever.

Both approaches reached **R² above 0.98** on independent validation, so the reported advantage is search efficiency rather than predictive quality. All studies were simulated; reduced wet-lab effort is an extrapolation, not a demonstrated outcome.
Connected Context · Feed7 Judgment

DASyR-LLM provides measured evidence that an LLM can accelerate a deterministic scientific search loop by critiquing and replacing candidate equations without improving final predictive fit. It reinforces architectures that reserve calculation and validation for specialized systems, while narrowing the claimed benefit to simulated search efficiency rather than real experimental savings.

How Kepler Built Verifiable AI for Financial Services — Vinoo GaneshBoth separate model judgment from deterministic numerical work: DASyR-LLM lets the LLM propose equation structures while symbolic regression evaluates them, paralleling Kepler’s boundary between agent choices and replayable calculations.Semantic Blindness: 500,000 Sensors Confused an LLM - Raahul Singh & Vanč Levstik, PhaidraBoth use the LLM to formulate domain-aware candidates or plans and deterministic machinery to resolve or test them, reducing the need to place the full search problem inside model context.Why Off-the-Shelf AI Doesn't Understand Money — Udi Menkes, IntuitIntuit argues that domain context alone is insufficient without verified outcomes; DASyR-LLM supplies a closed feedback loop in which chemically informed proposals are checked by numerical search and independent validation.Stop AI Agent Hallucinations: 5 Techniques + Production Patterns - Elizabeth Fuentes, AWSDASyR-LLM concretely instantiates the recommended pattern of structured data, validation, and runtime correction by alternating LLM critique with quantitatively evaluated symbolic candidates.
Context Map
agentresearchdata#harness-engineering#tool-use
Uncertainty
Both approaches reached **R² above 0.98** on independent validation, so the reported advantage is search efficiency rather than predictive quality. All studies were simulated; reduced wet-lab effort is an extrapolation, not a demonstrated outcome.