AI RESEARCH
Miller-Index-Based Latent Crystallographic Fracture Plane Reasoning with Vision-Language Models
arXiv CS.LG
•
ArXi:2605.20416v1 Announce Type: new We study whether multimodal large language models (MLLMs) can leverage crystallographic plane indices (Miller indices) as a structured latent representation for reasoning about fracture geometry.