mlsanityData to 7 Oct 2026

Contrast transfer functions help quantify neural network out-of-distribution generalization in HRTEM

arXiv 2512.09067 ↗First posted 9 Dec 2025cs.LG cond-mat.mtrl-scicited by 0

Not on the map. Why?

Abstract (arXiv)

Neural networks, while effective for tackling many challenging scientific tasks, are not known to perform well out-of-distribution (OOD), i.e., within domains which differ from their training data. Understanding neural network OOD generalization is paramount to their successful deployment in experimental workflows, especially when ground-truth knowledge about the experiment is hard to establish or experimental conditions significantly vary. With inherent access to ground-truth information and fine-grained control of underlying distributions, simulation-based data curation facilitates precise investigation of OOD generalization behavior. Here, we probe generalization with respect to imaging conditions of neural network segmentation models for high-resolution transmission electron microscopy (HRTEM) imaging of nanoparticles, training and measuring the OOD generalization of over 12,000 neural networks using synthetic data generated via random structure sampling and multislice simulation. Using the HRTEM contrast transfer function, we further develop a framework to compare information content of HRTEM datasets and quantify OOD domain shifts. We demonstrate that neural network segmentation models enjoy significant performance stability, but will smoothly and predictably worsen as imaging conditions shift from the training distribution. Lastly, we consider limitations of our approach in explaining other OOD shifts, such as of the atomic structures, and discuss complementary techniques for understanding generalization in such settings.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 34 of its 78 references that are in the map. First posted on arXiv 9 Dec 2025.

Builds on

34 of 78 references.

References · 34

  1. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 29,915
  2. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,604
  3. 2023Segment Anythingcited by 8,158
  4. 2019Invariant Risk Minimizationcited by 1,832
  5. 2018A Simple Unified Framework for Detecting Out-of-Distribution Samples and Adversarial Attackscited by 1,472
  6. 2021Towards Out-Of-Distribution Generalization: A Surveycited by 431
  7. 2019PyHessian: Neural Networks Through the Lens of the Hessiancited by 291
  8. 2021Accuracy on the Line: On the Strong Correlation Between Out-of-Distribution and In-Distribution Generalizationcited by 248
  9. 2020Domain Generalization using Causal Matchingcited by 224
  10. 2021Invariance Principle Meets Information Bottleneck for Out-of-Distribution Generalizationcited by 220
  11. 2021Vision Transformers are Robust Learnerscited by 208
  12. 2021Fishr: Invariant Gradient Variances for Out-of-Distribution Generalizationcited by 186
22 more
  1. 2022Diverse Weight Averaging for Out-of-Distribution Generalizationcited by 148
  2. 2021Towards a Theoretical Framework of Out-of-Distribution Generalizationcited by 94
  3. 2021OoD-Bench: Quantifying and Understanding Two Dimensions of Out-of-Distribution Generalizationcited by 86
  4. 2021Sketching Curvature for Efficient Out-of-Distribution Detection for Deep Neural Networkscited by 32
  5. 2021A statistical framework for efficient out of distribution detection in deep neural networkscited by 27
  6. 2020Review: Deep Learning in Electron Microscopycited by 20
  7. 2024A Survey on Evaluation of Out-of-Distribution Generalizationcited by 20
  8. 2022Visualizing high-dimensional loss landscapes with Hessian directionscited by 16
  9. 2020Machine Learning Pipeline for Segmentation and Defect Identification from High Resolution Transmission Electron Microscopy Datacited by 14
  10. 2024LCA-on-the-Line: Benchmarking Out-of-Distribution Generalization with Class Taxonomiescited by 12
  11. 2022An Empirical Study on Distribution Shift Robustness From the Perspective of Pre-Training and Data Augmentationcited by 11
  12. 2020Deep Learning Segmentation of Complex Features in Atomic-Resolution Phase Contrast Transmission Electron Microscopy Imagescited by 10
  13. 2024Model-free quantification of completeness, uncertainties, and outliers in atomistic machine learning using information theorycited by 10
  14. 2022Understanding the Influence of Receptive Field and Network Complexity in Neural-Network-Guided TEM Image Analysiscited by 9
  15. 2021Developing and Evaluating Deep Neural Network-based Denoising for Nanoparticle TEM Images with Ultra-low Signal-to-Noisecited by 8
  16. 2023Leveraging generative adversarial networks to create realistic scanning transmission electron microscopy imagescited by 8
  17. 2022Disentangling multiple scattering with deep learning: application to strain mapping from electron diffraction patternscited by 7
  18. 2024Ask Your Distribution Shift if Pre-Training is Right for Youcited by 7
  19. 2024SAM-I-Am: Semantic Boosting for Zero-shot Atomic-Scale Electron Micrograph Segmentationcited by 5
  20. 2023A robust synthetic data generation framework for machine learning in High-Resolution Transmission Electron Microscopy (HRTEM)cited by 3
  21. 2025Improving Multislice Electron Ptychography with a Generative Priorcited by 2
  22. 2025Generalized Few-Shot Out-of-Distribution Detectioncited by 1

Built on by

Not cited yet.