Find a story
Search Spins
Search titles, summaries, and missing voices across published articles — press releases, announcements, and media coverage.
3 results for “contrastive learning”
Contrastive Learning for Interpretable Anomaly Detection at Collider Experiments
Researchers introduced ORCA, a two-stage contrastive learning framework for anomaly detection in collider physics that improves sensitivity to new physics signals and enables interpretable attribution of anomalies to known physics processes using embedding geometry.
Aug 17, 2026
Beyond Decision Boundaries: Relational Geometry Attacks on Contrastive Embedding Manifolds
Researchers introduced a new adversarial attack framework that corrupts the relational geometry of contrastive embedding manifolds—targeting similarity structure rather than classification decisions—and demonstrated severe performance degradation on verification systems like Markmatch.
Aug 12, 2026
Breaking Structural Isolation: Scalable Graph Clustering via Community-Aware Sampling and Structural Entropy
A new unsupervised graph clustering framework called SCISE is introduced to address 'structural isolation' in mini-batch training by combining community-aware sampling and structural entropy constraints, showing improved performance on six benchmark datasets.
Jul 9, 2026