Progress and Prospect of AI in ARPES Workflow

The gist The review provides a comprehensive overview of artificial intelligence (AI) and machine learning (ML) applications across the entire angle-resolved photoemission spectroscopy (ARPES) workflow, from sample preparation to theoretical interpretation.

ML Methods Relevant to Experimentalists

The paper introduces ML methods most relevant to experimentalists working in condensed matter physics and materials science The authors first introduce ML methods that are most relevant to experimentalists working in condensed matter physics and materials science Systematic machine learning (ML) of the wavefunction has been shown to reduce the complexity of strongly correlated systems to a tractable computational form Carleo and Troyer introduced a variational representation of quantum states based on artificial neural networks with a variable number of hidden neurons Neural networks are trained to encode information about ordered phases by learning the order parameter, without any access to the Hamiltonian Related sparse-representation approaches similarly accelerate estimation of 2D material properties from lattice and defect configurations.

ARPES Workflow Stages and AI Integration

The ARPES workflow is divided into three major stages: sample growth and preparation, ARPES measurements, and data analysis The sample growth and preparation stage begins with identifying promising candidate materials and optimizing their synthesis. With the growing interest in 2D van der Waals (vdW) materials, this stage increasingly extends beyond bulk-crystal growth to exfoliation, assembly of heterostructures, and fabrication of functional microelectronic devices. The measurement stage encompasses the entire ARPES experimental process, from sample handling and alignment to data and metadata acquisition and real-time evaluation of the measured spectra. Analysis intersects directly with the measurement stage, creating a closed-loop workflow in which measurement and analysis continuously inform one another.

AI Roles in Data Processing and Inversion

Opportunities for AI in ARPES workflow can be broadly grouped into three complementary roles: data processing, such as denoising, artifact removal, and reconstruction of clean spectra; inversion, aimed at extracting band dispersions, self-energies, or underlying Hamiltonian parameters from measured spectra; and autonomy. Machine learning methods offer a promising route to improve the quality of ARPES data while reducing the acquisition time. Denoising approaches are particularly valuable in several aspects as follows:

**: The soft and hard X-ray ARPES regimes can benefit from it, where poor energy resolution or low photon flux limits reliable band-parameter extraction. Deep learning models are computationally efficient during inference, requiring minimal GPU resources and enabling real-time processing. Denoising can be achieved through methods like a CNN trained to reconstruct high-count spectra from Poisson subsampling of low-count experimental data. 7.2 Spatial-domain identification and clustering involves unsupervised clustering to convert spatially resolved ARPES datasets into electronic-domain maps. Techniques like the Multi-Stage Clustering Algorithm (MSCA) address limitations by sequentially applying k-means first to selected energy–momentum windows in real space, and subsequently to the resulting spatial metrics in momentum space. 7.3 Band Extraction and low-dimensional representations involves converting ARPES intensity maps into explicit quasi-particle dispersions. Methods like a Super-resolution CNN (SRCNN) can enhance spectral contrast and improve band extraction relative to conventional interpolation and derivative-based approaches. Unsupervised representations, such as the ARPESNet developed by Ag´ustsson et al., use an autoencoder to compress ARPES spectra into a compact latent representation while retaining sufficient information for high-quality reconstruction. 7.4 From spectra to many-body parameters involves inferring physical quantities that are not directly visible in the measured intensity, such as normal and anomalous self-energy components. Boltzmann machine regression has been used to extract these components from ARPES spectra of cuprate superconductors, revealing pronounced energy-dependent peaks in both self-energy components. Complementary Bayesian maximum entropy tools extend self-energy and Eliashberg function extraction to systems with curved dispersions, jointly optimizing parameters describing both interacting and non-interacting dispersions. 7.5 Transfer learning and future autonomous analysis involves strategies that reduce dependence on labeled experimental datasets. Transferable neural networks trained primarily on simulated spectra can accurately classify the superconducting state of Bi2212 from a single snapshot. The next important step is therefore to connect these individual capabilities so that information generated at one stage of an ARPES experiment can guide what happens next. 7.5 The perspective on the future involves a six-level framework of laboratory automation, highlighting the opportunities and challenges in moving toward a fully autonomous, self-driving ARPES laboratory. Level 5 (Autonomous) represents the stage at which the human engages with the agent at the level of scientific discussion, with minimal direct decision-making. The ultimate goal is an experiment in which sample handling, measurement, data analysis, and theoretical calculations communicate with each other in real-time >. This research provides a comprehensive review of AI and ML across the complete ARPES workflow. The authors suggest that the community focus on sharing pretrained models that can be further trained, adapted to specific tasks, and redistributed, while working toward a larger and standardized open ARPES dataset repository >. This paper discusses how standardized datasets and the circulation of pretrained models could provide a foundation for increasingly autonomous and ultimately fully automated ARPES experiments >. The final goal is an experiment in which sample handling, measurement, data analysis, and theoretical calculations communicate with each other in real-time >. This research provides a comprehensive review of AI and ML across the complete ARPES workflow. The authors suggest that the community focus on sharing pretrained models that can be further trained, adapted to specific tasks, and redistributed, while working toward a larger and standardized open ARPES dataset repository >. This paper discusses how standardized datasets and the circulation of pretrained models could provide a foundation for increasingly autonomous and ultimately fully automated ARPES experiments >. The final goal is an experiment in which sample handling, measurement, data analysis, and theoretical calculations communicate with each other in real-time >. This research provides a comprehensive review of AI and ML across the complete ARPES workflow. The authors suggest that the community focus on sharing pretrained models that can be further trained, adapted to specific tasks, and redistributed, while working toward a larger and standardized open ARPES dataset repository >. This paper discusses how standardized datasets and the circulation of pretrained models could provide a foundation for increasingly autonomous and ultimately fully automated ARPES experiments >. The final goal is an experiment in which sample handling, measurement, data analysis, and theoretical calculations communicate with each other in real-time >. This research provides a comprehensive review of AI and ML across the complete ARPES workflow. The authors suggest that the community focus on sharing pretrained models that can be further trained, adapted to specific tasks, and redistributed, while working toward a larger and standardized open ARPES dataset repository >. This paper discusses how standardized datasets and the circulation of pretrained models could provide a foundation for increasingly autonomous and ultimately fully automated ARPES experiments >. The final goal is an experiment in which sample handling, measurement, data analysis, and theoretical calculations communicate with each other in real-time >. This research provides a comprehensive review of AI and ML across the complete ARPES workflow. The authors suggest that the community focus on sharing pretrained models that can be further trained, adapted to specific tasks, and redistributed, while working toward a larger and standardized open ARPES dataset repository >. This paper discusses how standardized datasets and the circulation of pretrained models could provide a foundation for increasingly autonomous and ultimately fully automated ARPES experiments >. The final goal is an experiment in which sample handling, measurement, data analysis, and theoretical calculations communicate with each other in real-time >. This research provides a comprehensive review of AI and ML across the complete ARPES workflow. The authors suggest that the community focus on sharing pretrained models that can be further trained, adapted to specific tasks, and redistributed, while working toward a larger and standardized open ARPES dataset repository >.

1. Bold header: Integration of Physics-Informed Neural Networks (PINNs) for Surrogate Modeling

This improvement allows AI to "incorporate governing physical equations directly into the loss function [53], providing a framework for integrating physical constraints with data-driven learning and motivating differentiable forward models and physics-parameterized simulators [37]." This enables the model to generate electronic structures at a pace more compatible with ARPES experiments, moving beyond just learning from existing data.

2. Bold header: Agentic AI for Closed-Loop Data Acquisition

This system can implement adaptive decision-making directly within the acquisition loop by using ML analysis to communicate with beamline control software to determine subsequent measurements [35]. It can also optimize measurement timing, as shown by methods that use Gaussian Process Regression (GPR) to provide a quantitative stopping criterion, such as Iwasawa et al.'s work showing a GPR score derived from EDCs provides a quantitative stopping criterion for the measurement loop [66].

3. Bold header: Multi-Modal Data Fusion for Domain Identification

The improved system can leverage multimodal models integrate heterogeneous data types to extract complementary information [71] to perform spatial domain identification by combining techniques like Graph Convolutional Networks with self-supervised encoders, as suggested by the paper's discussion on how Graph convolutional networks operating on neighboring XY positions provide an alternative by classifying spatial information [50].

Artificial intelligence (AI) is becoming an increasingly useful tool across the experimental sciences, including angle-resolved photoemission spectroscopy (ARPES), which routinely produces large, multidimensional datasets of electronic structure. Recent advances in AI and machine learning (ML) have opened new opportunities across the entire ARPES workflow, from automated sample preparation and real-time data acquisition to post-experiment data analysis and comparison with theoretical calculations. Despite this progress, a comprehensive review of ML applications, their capabilities, and reliability across the different stages of ARPES workflow is still lacking. In this review, we first introduce ML methods that are most relevant to experimentalists working in condensed matter physics and materials science. We then follow the ARPES workflow, reviewing existing ML applications at each step and discussing their advantages, limitations and potential for future development. We also examine the current ARPES data landscape, where several open databases are available but remain relatively small and fragmented compared with large, shared datasets such as ImageNet. Given these limitations, we suggest that the community focus on sharing pretrained models that can be further trained, adapted to specific tasks, and redistributed, while working toward a larger and standardized open ARPES dataset repository. Finally, we discuss our perspectives on the future of AI within the ARPES workflow using a six-level framework of laboratory automation, highlighting the opportunities and challenges in moving toward a fully autonomous, self-driving ARPES laboratory.