Beyond the Final Label: Exploiting the Untapped Potential of Classification Histories in Astronomical Light Curve Analysis

2026-04-26

AI Summary

This work addresses the limitations of existing astronomical light curve classification methods, which rely solely on final labels and neglect the temporal evolution of classification decisions, thereby failing to meet the demands for early, stable, and high-accuracy object identification. To overcome this, we propose a novel framework that systematically leverages the temporal dynamics of classification history by modeling sequences of prediction probabilities using recurrent neural networks enhanced with additive attention mechanisms. Furthermore, we introduce a Wasserstein distance–based evaluation metric to holistically assess classifier performance in terms of accuracy, stability, and early recognition capability under limited observational data. Experiments on the ELAsTiCC challenge dataset demonstrate that our approach significantly improves overall classification performance and achieves a more balanced precision–recall trade-off, validating the effectiveness of both the proposed architecture and the new evaluation paradigm.

Technology Category

• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Reasoning under Uncertainty: Relational Probabilistic Models
• Natural Language Processing: Interpretability, Analysis, and Evaluation of NLP Models

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs
• Web Mining and Content Analysis: Models for Web evolution

Abstract

The Legacy Survey of Space and Time (LSST) on the Vera C. Rubin Observatory will generate a massive collection of time series (light curves) of the measured flux of transient and variable astronomical objects. With each new flux observation, light curve classifiers need to generate updated probability distributions over candidate classes, which will then be shared with the global community for the purpose of identifying interesting targets for follow-up observations as well as less time-sensitive analysis applications. Using the synthetic light curves and classification results of participating classifiers from the Extended LSST Astronomical Time-series Classification Challenge (ELAsTiCC), we investigate a novel framework to enhance existing light curve classifications by incorporating their classification histories and the temporal evolution of these histories. To demonstrate the potential of this approach, we introduce a model that combines a recurrent neural network and an additive attention module, which shows improved classification accuracy and more balanced precision-recall performance compared to existing classifiers from the challenge. Furthermore, at this stage, most, if not all, of the existing classifiers are evaluated by their final classification results on complete light curves; we propose new metrics that evaluate the stability, accuracy, and early classification performance of a classifier's predictions when using limited data by considering the Wasserstein distance between the temporally evolving classification probability distributions. Our metrics offer a more comprehensive perspective for model assessment by supplementing classical methods such as the confusion matrix and precision-recall.

Problem

• light curve classification
• classification history
• time-series analysis
• early classification
• astronomical transients

Innovation

• classification history
• recurrent neural network
• additive attention
• Wasserstein distance
• early classification