Vision language models struggle with physics aware photonic fiber design
PCFBench: How Far Can Large Vision-Language Models Go in Physics-aware Photonic Inverse Design?
Computer Vision and Pattern Recognition
Summary
Designing special optical fibers called photonic crystal fibers is usually hard and needs expert knowledge. The authors tested vision-language AI models to see if they can understand images of these fibers and help design new ones. They built a big benchmark called PCFBench with many tasks to check how well these models perceive, reason, and decide about fiber design. They found that while AI can recognize fiber shapes somewhat, it cannot use that understanding to make physics-based design decisions. Simply connecting different AI agents doesn't work well because early mistakes worsen later in the process.
What this means in practice
- •For fiber optics engineers: Use PCFBench to test AI tools for automating photonic crystal fiber design workflows combining image and physics understanding.
- •For machine learning developers: Evaluate and improve multi-task vision-language models on physics-grounded design challenges using PCFBench’s extensive dataset.
Authors
Shengchao Chen, Ting Shu, Sufen Ren
Abstract
Photonic crystal fibers (PCFs) underpin critical applications in telecommunications, sensing, and high-power laser delivery, yet their design remains a manual, expert-intensive process that couples geometric intuition with deep waveguide physics. Vision-language models (VLMs) offer a promising path toward automating PCF design, but can they truly reason about photonic physics from microscopy images, or do they merely recognize visual patterns? We introduce PCFBench, the first comprehensive benchmark for evaluating VLMs on the perception, reasoning, and decision steps of the photonic inverse-design loop. PCFBench spans 26 tasks organized into six categories of increasing cognitive demand (Geometry Perception, Physics Understanding, Text Generation, Multi-modal Reasoning, Inverse-Design Decision, and Code Generation), built from over 220K FDTD simulation samples across 18 structurally diverse fiber families. We evaluate both individual VLMs on isolated tasks and multi-agent systems (MAS) on the full sequential pipeline from perception to design. Our results reveal a striking perception-design gap: current models can identify fiber geometry with moderate success, yet consistently fail to translate that visual understanding into physics-aware design actions. Moreover, cascading agents through the pipeline exposes significant error propagation, where upstream perception mistakes compound into downstream design failures. These findings suggest that neither model scaling nor naive agent chaining suffices, and that closing the loop demands explicit physical grounding. PCFBench provides the community with a rigorous, multi-dimensional testbed to measure progress toward AI-assisted photonic engineering. The project is available at https://pcfbench.github.io/.