Research map / Large language models / Agents, code and math
CoEvo: Continual Evolution of Symbolic Solutions Using Large Language Models
Ping Guo, Qingfu Zhang, Xi Lin
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 changed | Where | Measured | Effect |
|---|---|---|---|
| use CoEvo (idea tree, multiple representations, knowledge library) instead of LLM-SR | gpt-3.5-turbo backbone · Oscillation 2 (OOD) | NMSE: 3.03e-3 -> 1.32e-10 | Improved |
| use CoEvo instead of LLM-SR | gpt-3.5-turbo backbone · E. coli growth (ID) | NMSE: 0.0207 -> 1.58e-9 | Improved |
| use CoEvo instead of LLM-SR | gpt-4o-mini backbone · Oscillation 2 (OOD) | NMSE: 3.11e-5 -> 2.21e-4 | Improved |
| use CoEvo instead of LLM-SR | gpt-4o-mini backbone · Oscillation 1 (ID) | NMSE: 5.14e-9 -> 1.28e-8 | Hurt |
| use CoEvo instead of LLM-SR | gpt-3.5-turbo backbone · Stress-Strain (ID) | NMSE: 0.0017 -> 0.0020 | No clear effect |
| swap backbone from gpt-3.5-turbo to gpt-4o-mini | CoEvo · Oscillation 2 (OOD) | NMSE: 1.32e-10 -> 2.21e-4 | Hurt |
| add idea knowledge library when generating new solutions (ablation) | CoEvo · Oscillation 1 and E. coli growth | NMSE | Improved |
| add knowledge library (ablation) | CoEvo · Stress-Strain | NMSE | No clear effect |
| use CoEvo instead of LLM-SR on solution validity | CoEvo · all four benchmarks | valid solution rate: higher valid solution ratio | Improved |
| use stronger LLM (gpt-4o-mini) for valid solution ratio | CoEvo · Stress-Strain | valid solution rate | Hurt |
Closest papers by meaning
- Evolution Fine-Tuning: Learning to Discover Across 371 Optimization Tasks (NeurIPS 2026 · similarity 0.87)
- Automated Creation of Reusable and Diverse Toolsets for Enhancing LLM Reasoning (AAAI 2025 · similarity 0.87)
- FunctionEvolve: Structure-Guided Symbolic Regression with LLMs (NeurIPS 2026 · similarity 0.87)
- LLM-SR: Scientific Equation Discovery via Programming with Large Language Models (ICLR 2025 · similarity 0.87)
- LLM-SRBench: A New Benchmark for Scientific Equation Discovery with Large Language Models (ICML 2025 · similarity 0.86)
- DiscoveryBench: Towards Data-Driven Discovery with Large Language Models (ICLR 2025 · similarity 0.86)
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.
