![[2506.05325] Quasiparticle Interference Kernel Extraction with Variational Autoencoders via Latent Alignment](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2506.05325] Quasiparticle Interference Kernel Extraction with Variational Autoencoders via Latent Alignment
Summary
This article presents an AI-based framework for extracting quasiparticle interference (QPI) kernels from complex scattering images, improving accuracy and generalization in quantum material analysis.
Why It Matters
The research addresses a significant challenge in quantum material studies, where traditional methods struggle with complex scattering conditions. By leveraging AI, this framework enhances the extraction of critical data, potentially advancing material science and related fields.
Key Takeaways
- Introduces an AI framework for QPI kernel extraction.
- Utilizes a two-step learning strategy for improved accuracy.
- Demonstrates effectiveness on both synthetic and real QPI data.
- Constructs a diverse dataset of 100 unique kernels for training.
- Achieves better generalization to unseen kernels compared to traditional methods.
Computer Science > Machine Learning arXiv:2506.05325 (cs) [Submitted on 5 Jun 2025 (v1), last revised 13 Feb 2026 (this version, v3)] Title:Quasiparticle Interference Kernel Extraction with Variational Autoencoders via Latent Alignment Authors:Yingshuai Ji, Haomin Zhuang, Matthew Toole, James McKenzie, Xiaolong Liu, Xiangliang Zhang View a PDF of the paper titled Quasiparticle Interference Kernel Extraction with Variational Autoencoders via Latent Alignment, by Yingshuai Ji and 4 other authors View PDF HTML (experimental) 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, because many different kernels can combine to produce almost the same observed image, and noise or overlaps further obscure the true signal. Existing solutions to this extraction problem rely on manually zooming into small local regions with isolated single-scatterers. This is infeasible for real cases where scattering conditions are too complex. In this work, we propose the first AI-based framework for QPI kernel extraction, which models the space of physically valid kernels and uses this knowledge to guide the inverse mapping. We introduce a two-step learning strategy that decouples kernel representation learning from observation-to-kernel inference. In the first step, we train a variat...