Reproduce “DASyR-LLM: Domain-Aware Symbolic Regression with LLMs for Kinetic Model Discovery” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Roberto Aliaga Medina, Paulina Quintanilla, Antonio del Rio Chanona
An iterative symbolic regression framework, DASyR-LLM, incorporates large language models to inject domain knowledge into kinetic model discovery, leading to a substantial reduction in experimental iterations (41.7% to 79.3%) while maintaining high predictive accuracy and mechanistic plausibility. This approach accelerates the identification of interpretable kinetic models by leveraging LLM-driven critique and proposal generation.
Key insights
- Large Language Models can effectively act as both qualitative critics and generative proposers of kinetic model hypotheses, injecting crucial domain knowledge into data-driven discovery.
- The synergistic integration of data-driven symbolic regression with knowledge-driven LLM guidance and model-based design of experiments significantly enhances model discovery efficiency.
- LLMs can identify and correct structural limitations in statistically strong SR-generated models, leading to mechanistically more plausible and interpretable results.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.