Model-Agnostic Signal Discovery with Machine Learning: Bridging the Gap Between Theory and Practice
Not on the map. Why?
Abstract (arXiv)
Searches for new phenomena in complex scientific data are predominantly model-dependent, optimized for specific hypotheses, and therefore limited in their coverage of the space of possible signals. Recently, new AI-based model-agnostic search strategies, many of which have been pioneered in high-energy physics, have been proposed which provide a complementary paradigm, prioritizing broad exploration over tailored analyses. These techniques offer an opportunity to enhance the overall discovery potential of modern experiments, especially in regimes where theoretical guidance is scarce. In this document, we review the conceptual framework behind the main classes of AI-based model-agnostic strategies. We discuss the potential pitfalls of these methods, and strategies for their validation and interpretation. We aim for this document to serve as a useful reference both for practitioners and for researchers interested in learning more about these model-agnostic search strategies.
Key facts. Cited by 3 papers in the map (data through 14 Oct 2026). Builds on 29 of its 103 references that are in the map. First posted on arXiv 29 May 2026.
Builds on
29 of 103 references.
References · 29
- 2015Variational Inference with Normalizing Flowscited by 2,170
- 2016Revisiting Classifier Two-Sample Testscited by 268
- 2020Why Normalizing Flows Fail to Detect Out-of-Distribution Datacited by 221
- 2019Input complexity and out-of-distribution detection with likelihood-based generative modelscited by 186
- 2015On Wasserstein Two Sample Testing and Related Families of Nonparametric Testscited by 166
- 2021MMD Aggregated Two-Sample Testcited by 75
- 2020Anomaly Detection with Density Estimationcited by 72
- 2017Classification without labels: Learning from mixed samples in high energy physicscited by 43
- 2016Classification accuracy as a proxy for two sample testingcited by 42
- 2023MMD-FUSE: Learning and Combining Kernels for Two-Sample Testing Without Data Splittingcited by 33
- 2023Machine Learning for Anomaly Detection in Particle Physicscited by 30
- 2021The Dark Machines Anomaly Score Challenge: Benchmark Data and Model Independent Event Classification for the Large Hadron Collidercited by 29
17 more
- 2020Simulation Assisted Likelihood-free Anomaly Detectioncited by 27
- 2022Learning new physics efficiently with nonparametric methodscited by 21
- 2020Adversarially Learned Anomaly Detection on CMS Open Data: re-discovering the top quarkcited by 18
- 2025Out-of-Distribution Detection Methods Answer the Wrong Questionscited by 14
- 2020Simulation-Assisted Decorrelation for Resonant Anomaly Detectioncited by 13
- 2020Unsupervised in-distribution anomaly detection of new physics through conditional density estimationcited by 12
- 2023High-dimensional and Permutation Invariant Anomaly Detectioncited by 12
- 2023CURTAINs Flows For Flows: Constructing Unobserved Regions with Maximum Likelihood Estimationcited by 10
- 2023Goodness of fit by Neyman-Pearson testingcited by 10
- 2024Refereeing the Referees: Evaluating Two-Sample Tests for Validating Generators in Precision Sciencescited by 9
- 2023Fast kernel methods for Data Quality Monitoring as a goodness-of-fit testcited by 8
- 2022Null Hypothesis Test for Anomaly Detectioncited by 7
- 2024Multiple testing for signal-agnostic searches of new physics with machine learningcited by 7
- 2022Resonant Anomaly Detection with Multiple Reference Datasetscited by 4
- 2025Comparing Generative Models with the New Physics Learning Machinecited by 4
- 2024SIGMA: Single Interpolated Generative Model for Anomaliescited by 2
- 2025Minimax-Optimal Two-Sample Test with Sliced Wassersteincited by 1
Built on by
Citing papers per year: 2026: 3.
3 papers.