CYGNO Implements ML-Based Event Reconstruction and Selection
An arXiv preprint (submitted 28 Jan 2026) reports two machine-learning pipelines developed for the CYGNO optical-readout time-projection chamber to improve real-time triggering, data reduction, and background discrimination. A convolutional autoencoder trained on pedestal images enables unsupervised ROI extraction, retaining 93.0±0.2% of reconstructed signal while discarding 97.8±0.1% of image area with ~25 ms per-frame inference on a consumer GPU; a weakly supervised CWoLa classifier applied to AmBe data isolates nuclear-recoil-like, compact circular topologies approaching the mixture-imposed discrimination limit.
Scoring Rationale
Strong practical performance and actionability demonstrated on prototype data; limited by single-experiment arXiv preprint status.
Practice interview problems based on real data
1,500+ SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
