Vision Transformer for Transient Noise Classification

2025-10-06

AI Summary

Gravitational-wave detection by LIGO is severely hindered by transient non-Gaussian noise artifacts (“glitches”), necessitating high-accuracy, multi-class classification. This work introduces the first application of a pre-trained Vision Transformer (ViT-B/32) to classify glitches in time-frequency representations of gravitational-wave data, unifying Gravity Spy’s original 22 classes with two newly identified classes from the O3a observing run. By jointly training on the augmented dataset and fine-tuning ViT, we achieve 92.26% classification accuracy—substantially outperforming conventional CNN-based approaches. Our method establishes a scalable, robust deep learning framework for real-time glitch identification and mitigation during O3 and future observing runs. Beyond performance gains, this study extends the applicability of vision transformers to astrophysical time-series signal analysis, demonstrating their efficacy in modeling complex, structured noise patterns in gravitational-wave data.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Security and Privacy: Data transparency and provenance
• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

Transient noise (glitches) in LIGO data hinders the detection of gravitational waves (GW). The Gravity Spy project has categorized these noise events into various classes. With the O3 run, there is the inclusion of two additional noise classes and thus a need to train new models for effective classification. We aim to classify glitches in LIGO data into 22 existing classes from the first run plus 2 additional noise classes from O3a using the Vision Transformer (ViT) model. We train a pre-trained Vision Transformer (ViT-B/32) model on a combined dataset consisting of the Gravity Spy dataset with the additional two classes from the LIGO O3a run. We achieve a classification efficiency of 92.26%, demonstrating the potential of Vision Transformer to improve the accuracy of gravitational wave detection by effectively distinguishing transient noise. Key words: gravitational waves --vision transformer --machine learning

Problem

• Classifying transient noise in LIGO data using Vision Transformer
• Improving gravitational wave detection accuracy through glitch classification
• Expanding classification to 24 noise categories from LIGO observations

Innovation

• Vision Transformer classifies LIGO transient noise
• Pre-trained ViT model trained on expanded dataset
• Achieves 92.26% classification accuracy for glitches