QuPAINT：用于量子材料表征的物理感知多模态推理
QuPAINT: Physics-Aware Multimodal Reasoning for Quantum Material Characterization
查看机构详情
- University of Arkansas（阿肯色大学）
- University of Utah（犹他大学）
机构由 AI 辅助整理，请以论文原文为准。
浏览论文内容
中文总结 AI 辅助
QuPAINT提出物理感知多模态框架，结合合成数据与物理信息注意力，解决量子材料光学表征的域迁移问题，并在基准上实现最先进性能。
中文摘要 AI 辅助
通过光学显微镜表征二维（2D）量子材料，需要定位剥离的薄片，并根据微妙的衬度光学对比和干涉色确定其层厚，以选择合适的薄片用于器件制造。然而，模型面临合成到真实的域偏移以及材料、衬底、实验室和成像条件之间的差异。我们提出了QuPAINT，一个用于可迁移量子薄片表征的物理感知多模态框架。合成材料框架（Synthia）生成多样化的合成显微图像，同时保留依赖于层的光学行为。利用这些图像，我们构建了QMat-Instruct，一个多模态指令数据集，其中包含基于已验证注释生成的图像特定推理轨迹，并约束于可观察的光学线索。QuPAINT通过物理信息注意力（PIA）整合这些信号，该机制将衬底相对的光学先验注入视觉表示中，以支持基于物理的多模态推理。为了评估，我们引入了QF-Bench，据我们所知，这是该问题最大的真实世界基准，涵盖多种显微和衬底条件。利用其已验证的注释，我们研究了计数、视觉定位、推理质量、置信度校准以及向未见材料的迁移。QuPAINT-8B大幅优于先前方法，并在一般和单层薄片检测方面均建立了最先进的性能。额外实验表明，基于图像的监督改善了严格的空间定位和置信度校准，同时保持了对未见材料的稳健的一般薄片检测。
英文摘要
Characterizing two-dimensional (2D) quantum materials by optical microscopy requires localizing exfoliated flakes and determining their layer thickness from subtle optical contrast and interference color to select suitable flakes for device fabrication. However, models face synthetic-to-real domain shifts and variation across materials, substrates, laboratories, and imaging conditions. We present QuPAINT, a physics-aware multimodal framework for transferable quantum flake characterization. The Synthetic Materials Framework (Synthia) generates diverse synthetic microscopy images while preserving layer-dependent optical behavior. Using these images, we construct QMat-Instruct, a multimodal instruction dataset with image-specific reasoning traces generated from verified annotations and constrained to observable optical cues. QuPAINT integrates these signals through Physics-Informed Attention (PIA), which injects substrate-relative optical priors into the visual representation to support grounded multimodal reasoning. For evaluation, we introduce QF-Bench, to our knowledge, the largest real-world benchmark for this problem, spanning diverse microscopy and substrate conditions. Using its verified annotations, we study counting, visual grounding, reasoning quality, confidence calibration, and transfer to an unseen material. QuPAINT-8B substantially outperforms prior methods and establishes state-of-the-art performance for both general and monolayer flake detection. Additional experiments show that image-grounded supervision improves strict spatial grounding and confidence calibration while preserving robust general flake detection on the unseen material.
