Adapting General-Purpose Foundation Models for X-ray Ptychography in Low-Data Regimes

TL;DR

This paper introduces PtychoBench, a benchmark for ptychographic analysis, comparing fine-tuning and in-context learning strategies for adapting foundation models to scientific microscopy tasks in low-data scenarios.

Contribution

The paper presents a new benchmark and systematically evaluates specialization strategies, revealing task-dependent optimal adaptation methods for scientific AI applications.

Findings

Supervised fine-tuning and ICL are complementary for visual artifact detection.

ICL on large models outperforms SFT for textual parameter recommendation.

Context-aware prompting enhances performance across tasks.

Abstract

The automation of workflows in advanced microscopy is a key goal where foundation models like Language Models (LLMs) and Vision-Language Models (VLMs) show great potential. However, adapting these general-purpose models for specialized scientific tasks is critical, and the optimal domain adaptation strategy is often unclear. To address this, we introduce PtychoBench, a new multi-modal, multi-task benchmark for ptychographic analysis. Using this benchmark, we systematically compare two specialization strategies: Supervised Fine-Tuning (SFT) and In-Context Learning (ICL). We evaluate these strategies on a visual artifact detection task with VLMs and a textual parameter recommendation task with LLMs in a data-scarce regime. Our findings reveal that the optimal specialization pathway is task-dependent. For the visual task, SFT and ICL are highly complementary, with a fine-tuned model guided by context-aware examples achieving the highest mean performance (Micro-F1 of 0.728). Conversely, for the textual task, ICL on a large base model is the superior strategy, reaching a peak Micro-F1 of 0.847 and outperforming a powerful "super-expert" SFT model (0-shot Micro-F1 of 0.839). We also confirm the superiority of context-aware prompting and identify a consistent contextual interference phenomenon in fine-tuned models. These results, benchmarked against strong baselines including GPT-4o and a DINOv3-based classifier, offer key observations for AI in science: the optimal specialization path in our benchmark is dependent on the task modality, offering a clear framework for developing more effective science-based agentic systems.