← Back to all news
arXiv Computer Vision By Luis Bauer, Annalena Kofler, Andr\'es Asensio Ramos, Aparna Venkataramanasastry, Valent\'in Mart\'inez Pillet, Bernhard Sch\"olkopf

ARENA: Active Region attENtion Aggregation for interpretable solar flare prediction

Read the original on arXiv Computer Vision →

The paper introduces ARENA, a neural network that predicts solar flares by aggregating information from multiple active‑region cutouts and learning to focus on the most relevant region. It uses weakly supervised multiple‑instance learning, requiring only a flare/no‑flare label, and its attention mechanism localizes the driving active region without explicit annotations. ARENA attains a Heidke skill score of 0.45, comparable to existing methods, showing that active‑region aggregation can yield efficient and interpretable flare predictions.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Computer Vision.

Related stories

arXiv Computer Vision
Sep 18

Semi-LAR: Semi-supervised Contrastive Learning with Linear Attention for Removal of Nighttime Flares

The paper introduces Semi-LAR, a semi‑supervised framework for removing nighttime lens flares. It uses an adaptive pseudo‑label repository that refines supervision through quality assessment, momentum updates, and invalid label filtering. A flare‑aware contrastive loss treats flare‑contaminated inputs as negatives, encouraging representations that distinguish flare patterns while aligning with reliable pseudo targets.

By Xiyu Zhu, Wei Wang, Kui Jiang, Zhengguo Li
benchmarks
More like this →
arXiv AI
Sep 24

PISCES: Physics-Informed Solar-wind Convolutional autoEncoder for Space-weather Anomaly Detection and Early Warning

PISCES is a physics‑informed convolutional autoencoder designed to detect solar‑wind transients for space‑weather early warning. Trained on OMNI solar‑wind data without catalog labels, its loss incorporates magnetic field consistency, temperature‑velocity relations, the Parker spiral angle, and temporal smoothness penalties. During inference, PISCES decomposes the anomaly score into magnetic, plasma, physics‑relation, and residual components, enabling alarms that can precede observed sudden commencements and positive sudden impulses.

By Kevin Lee, Alison J. March
More like this →