Quantum-Assisted Multimodal Anomaly Detection for Cross-Experiment New Physics Signals

While Schuhmacher et al. (2023) showed the promise of quantum support vector classifiers for anomaly detection in LHC data, their approach focused on artificial anomalies and single-experiment datasets. My proposal is to create a unified, quantum-accelerated anomaly detection framework that ingests data from multiple experimental domains—collider events (LHC), neutrino oscillations (reactor/short baseline), and astroparticle observations (e.g., IceCube, cosmic rays). This would enable the search for correlated or repeated anomalous patterns across very different experimental signatures, potentially revealing multi-channel evidence for BSM physics (such as dark sector particles or sterile neutrinos impacting both collider and cosmic-ray observations). The quantum component could address the curse of dimensionality and entanglement between features in this huge, heterogeneous data space. This is especially novel because it synthesizes techniques and data from multiple subfields, moving beyond the single-source focus of prior work, and could uncover BSM signals that would otherwise be missed in isolated analyses.

References:

1. Unravelling physics beyond the standard model with classical and quantum anomaly detection. Julian Schuhmacher, Laura Boggia, Vasilis Belis, E. Puljak, M. Grossi, M. Pierini, S. Vallecorsa, F. Tacchino, P. Barkoutsos, I. Tavernelli (2023). Machine Learning: Science and Technology.
2. Unravelling physics beyond the standard model with classical and quantum anomaly detection. Julian Schuhmacher, Laura Boggia, Vasilis Belis, E. Puljak, M. Grossi, M. Pierini, S. Vallecorsa, F. Tacchino, P. Barkoutsos, I. Tavernelli (2023). Machine Learning: Science and Technology.