SMCEvolve: Principled Scientific Discovery via Sequential Monte Carlo Evolution
Quick Answer
SMCEvolve introduces a principled framework for LLM-driven program evolution, enhancing automated scientific discovery.
Quick Take
SMCEvolve introduces a principled framework for LLM-driven program evolution, enhancing automated scientific discovery. By employing a Sequential Monte Carlo sampler, it achieves superior performance across benchmarks in math, algorithm efficiency, and symbolic regression, while minimizing LLM calls. The method ensures convergence and provides a finite-sample complexity analysis for effective resource management.
Key Points
- SMCEvolve uses a Sequential Monte Carlo sampler for program evolution.
- It outperforms state-of-the-art systems in multiple benchmarks.
- The framework includes adaptive resampling and automatic convergence control.
- A finite-sample complexity analysis is provided for resource budgeting.
- Fewer LLM calls are required for achieving target approximation errors.
Paper Resources
📖 Reader Mode
~2 min readAbstract:LLM-driven program evolution has emerged as a powerful tool for automated scientific discovery, yet existing frameworks offer no principled guide for designing their individual components and provide no guarantee that the search converges. We introduce SMCEvolve, which recasts program search as sampling from a reward-tilted target distribution and approximates it with a Sequential Monte Carlo (SMC) sampler. From this view, three core mechanisms emerge as principled components: adaptive parent resampling, mixture of mutation with acceptance, and automatic convergence control. We further provide a finite-sample complexity analysis that bounds the LLM-call budget required to reach a target approximation error. Across math, algorithm efficiency, symbolic regression, and end-to-end ML research benchmarks, SMCEvolve surpasses state-of-the-art evolving systems while using fewer LLM calls under self-determined termination. The code is available at this https URL.
| Subjects: | Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Multiagent Systems (cs.MA) |
| Cite as: | arXiv:2605.15308 [cs.AI] |
| (or arXiv:2605.15308v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2605.15308 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Jiachen Jiang [view email]
[v1]
Thu, 14 May 2026 18:21:08 UTC (4,977 KB)
— Originally published at arxiv.org
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.AI
See more →Procedural Memory Distillation: Online Reflection for Self-Improving Language Models
Procedural Memory Distillation (PMD) enhances reinforcement learning by converting cross-episode signals into reusable memory, improving Qwen3-8B and OLMo3-Instruct-7B models by 3.8-5.5% on SCIKNOWEVAL and 7.9-13.6% on . The co-evolution of policy and memory allows for more effective self-supervision, demonstrating significant performance gains when both components are active.