Autonomous Algorithm Discovery for Ptychography via Evolutionary LLM Reasoning

2026-03-05

AI Summary

This work addresses the limitations of handcrafted regularization functions in ptychographic imaging, which often constrain reconstruction quality. To overcome this, the authors propose Ptychi-Evolve, a novel framework that, for the first time, integrates large language models (LLMs) with semantic-guided evolutionary strategies to automatically discover and optimize new regularization algorithms. The approach leverages LLM-driven code generation, semantic-aware crossover and mutation operations, and lineage tracking of evolved algorithms to ensure interpretability and reproducibility. Evaluated on three challenging datasets, Ptychi-Evolve achieves significantly superior reconstruction performance compared to conventional methods, with improvements of up to 0.26 in SSIM and 8.3 dB in PSNR.

Technology Category

• Machine Learning: Evolutionary Learning
• Search and Optimization: Evolutionary Computation
• Computer Vision: Learning & Optimization for CV

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Ptychography is a computational imaging technique widely used for high-resolution materials characterization, but high-quality reconstructions often require the use of regularization functions that largely remain manually designed. We introduce Ptychi-Evolve, an autonomous framework that uses large language models (LLMs) to discover and evolve novel regularization algorithms. The framework combines LLM-driven code generation with evolutionary mechanisms, including semantically-guided crossover and mutation. Experiments on three challenging datasets (X-ray integrated circuits, low-dose electron microscopy of apoferritin, and multislice imaging with crosstalk artifacts) demonstrate that discovered regularizers outperform conventional reconstructions, achieving up to +0.26 SSIM and +8.3~dB PSNR improvements. Besides, Ptychi-Evolve records algorithm lineage and evolution metadata, enabling interpretable and reproducible analysis of discovered regularizers.

Problem

• Ptychography
• regularization
• algorithm discovery
• computational imaging
• autonomous optimization

Innovation

• autonomous algorithm discovery
• large language models
• evolutionary optimization
• ptychographic reconstruction
• regularization design