arXiv subjects

Andreas Werbrouck

Publications and source records attributed to Andreas Werbrouck.

4 recordsLinked to original sources

arXiv2026

Fast 4D-STEM-based phase mapping for amorphous and mixed materials

Interpretation and mapping strategies for 4D-scanning transmission electron microscopy (4D-STEM) are well-developed for crystalline materials, yet in the case of amorphous and mixed materials it is significantly more challenging to separate different phases. Non-negative matrix factorization (NMF) in principle would allow separation of 4D-STEM data into components with interpretable diffraction signatures and intensity maps, independent of the crystalline, amorphous or mixed nature of the material. However, adoption of NMF in this field is hampered by large datasets and conceptual hurdles: NMF tackles a non-convex optimization problem, requiring iterative algorithms. Additionally, the stopping condition has to be chosen carefully. In this work, we show that the factorization of large 4D-STEM datasets can be drastically accelerated using a QB decomposition (i.e. randomized NMF or RNMF), leading to much shorter time per iteration. This allows structure-independent phase mapping on very large 4D-STEM datasets. We validate this approach on a synthetic literature dataset (mixed ZrCuAl), before mapping a thin TiO$_2$ layer on top of SiO$_2$, and an interface between a lithium-ion cathode and solid state electrolyte. We also demonstrate that, before using NMF to transform the data on an interpretable, nonnegative basis, principal component analysis (PCA) can be used for fast exploratory analysis to assess dataset dimensionality and linearity.

cond-mat.mtrl-sci↗

arXiv2026

Using superpixels for interpretable feature reduction in large 2D diffraction datasets

Large 2D diffraction datasets, consisting of hundreds or thousands of measurements, are commonly acquired with 4D-STEM electron diffraction or at synchrotron X-ray beamlines. Machine learning and artificial intelligence offer great promise for analyzing these datasets. However, the sheer volume of data presents a significant data processing bottleneck. Cropping the detector and pixel binning are standard ways to reduce data size. Here we propose grouping and averaging pixels into superpixels of variable area. High-information regions are sampled densely, while low-information areas are collected into larger superpixels. In the process, symmetries in the data are captured and exploited, making this approach suitable for preprocessing 2D diffraction data. We compare two variance-minimizing methods: K-means clustering (top-down) and agglomerative clustering (bottom-up) and demonstrate superior scaling and interpretability for the bottom-up method. As these methods are distance-based, we demonstrate that the construction of superpixels can be accelerated using Gaussian random projection. Finally we show over 100-fold acceleration for phase mapping with Non-negative matrix factorization on a 4D-STEM dataset when superpixels are used as a preprocessing step.

cond-mat.mtrl-sci↗

arXiv2026

LLM Agents for Knowledge Discovery in Atomic Layer Processing

Large Language Models (LLMs) have garnered significant attention for several years now. Recently, their use as independently reasoning agents has been proposed. In this work, we test the potential of such agents for knowledge discovery in materials science. We repurpose LangGraph's tool functionality to supply agents with a black box function to interrogate. In contrast to process optimization or performing specific, user-defined tasks, knowledge discovery consists of freely exploring the system, posing and verifying statements about the behavior of this black box, with the sole objective of generating and verifying generalizable statements. We provide proof of concept for this approach through a children's parlor game, demonstrating the role of trial-and-error and persistence in knowledge discovery, and the strong path-dependence of results. We then apply the same strategy to show that LLM agents can explore, discover, and exploit diverse chemical interactions in an advanced Atomic Layer Processing reactor simulation using intentionally limited probe capabilities without explicit instructions.

cs.AI↗

arXiv2025

Active Learning and Explainable AI for Multi-Objective Optimization of Spin Coated Polymers

Spin coating polymer thin films to achieve specific mechanical properties is inherently a multi-objective optimization problem. We present a framework that integrates an active Pareto front learning algorithm (PyePAL) with visualization and explainable AI techniques to optimize processing parameters. PyePAL uses Gaussian process models to predict objective values (hardness and elasticity) from the design variables (spin speed, dilution, and polymer mixture), guiding the adaptive selection of samples toward promising regions of the design space. To enable interpretable insights into the high-dimensional design space, we utilize UMAP (Uniform Manifold Approximation and Projection) for two-dimensional visualization of the Pareto front exploration. Additionally, we incorporate fuzzy linguistic summaries, which translate the learned relationships between process parameters and performance objectives into linguistic statements, thus enhancing the explainability and understanding of the optimization results. Experimental results demonstrate that our method efficiently identifies promising polymer designs, while the visual and linguistic explanations facilitate expert-driven analysis and knowledge discovery.

cs.LG↗