Seeing the Invisible: Machine learning-Based QPI Kernel Extraction via Latent Alignment

2025-06-05

AI Summary

Robustly extracting the single-scattering kernel (i.e., the physical kernel) from quantitative phase imaging (QPI) data under multiple scattering is a severely ill-posed inverse problem. This paper introduces the first two-stage AI framework for this task: first, a physics-driven variational autoencoder (VAE) maps kernels to an interpretable latent space; second, latent-space alignment enables robust mapping from observed QPI images to the underlying kernel. The method integrates physics-constrained synthetic data generation with supervised contrastive encoding to ensure physical consistency and generalizability of solutions. Evaluated on a large-scale dataset comprising 100 distinct real-world physical kernels, our approach improves kernel reconstruction accuracy by 32% over end-to-end baselines and significantly enhances robustness in strongly entangled scattering regimes. It establishes a new paradigm for QPI inversion—interpretable, generalizable, and grounded in physical priors.

Technology Category

• Computer Vision: Computational Photography, Image & Video Synthesis
• Machine Learning: Kernel Methods
• Knowledge Representation and Reasoning: Qualitative Reasoning

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI
• Security and Privacy: Data transparency and provenance

Abstract

Quasiparticle interference (QPI) imaging is a powerful tool for probing electronic structures in quantum materials, but extracting the single-scatterer QPI pattern (i.e., the kernel) from a multi-scatterer image remains a fundamentally ill-posed inverse problem. In this work, we propose the first AI-based framework for QPI kernel extraction. We introduce a two-step learning strategy that decouples kernel representation learning from observation-to-kernel inference. In the first step, we train a variational autoencoder to learn a compact latent space of scattering kernels. In the second step, we align the latent representation of QPI observations with those of the pre-learned kernels using a dedicated encoder. This design enables the model to infer kernels robustly even under complex, entangled scattering conditions. We construct a diverse and physically realistic QPI dataset comprising 100 unique kernels and evaluate our method against a direct one-step baseline. Experimental results demonstrate that our approach achieves significantly higher extraction accuracy, and improved generalization to unseen kernels.

Problem

• Extracting single-scatterer QPI patterns from multi-scatterer images
• Solving ill-posed inverse problem in QPI kernel extraction
• Improving accuracy and generalization in QPI kernel inference

Innovation

• AI-based framework for QPI kernel extraction
• Two-step learning strategy decouples representation
• Variational autoencoder learns compact latent space