FALCO: a Foundation model of Astronomical Light Curves for time dOmain astronomy

Abstract

Time-domain surveys have advanced astronomical research by revealing diverse variable phenomena, from stellar flares to transient events. The scale and complexity of survey data, along with the demand for rapid classification, present significant challenges for analysis. While machine learning offers solutions, most existing models are tailored to single tasks, struggle to generalize, and depend heavily on large, accurately labeled datasets. We introduce FALCO, a foundation model for astronomical light curve analysis in time-domain astronomy. This work presents the initial version of FALCO trained via self-supervised learning on unlabeled Kepler light curves using a Transformer-based architecture. The model has been evaluated on three distinct tasks and demonstrates strong generalization: achieving 95 percent accuracy in stellar variability classification across eight classes, an overall RMSE of 0.1305 dex in surface gravity estimation (notably improved to below 0.08 dex when log g is less than 1, and approximately 0.02 dex near log g equals 3), and 87 percent precision in flare identification. These results highlight the model's versatility and ability to learn generalizable representations from light curves, enabling straightforward adaptation to diverse tasks. We further analyze the impact of model scaling and sequence length, finding performance improves with larger models and longer input sequences. We also apply FALCO to derive surface gravity (log g) measurements for 179,732 Kepler stars from their light curves.

References in corpus (49)

• SMOTE: Synthetic Minority Over-sampling Technique
• Array Programming with NumPy
• Astropy: A Community Python Package for Astronomy
• Observation of Gravitational Waves from a Binary Black Hole Merger
• The Sloan Digital Sky Survey: Technical Summary
• GW170817: Observation of Gravitational Waves from a Binary Neutron Star Inspiral
• The Astropy Project: Building an inclusive, open-science project and status of the v2.0 core package
• The Transiting Exoplanet Survey Satellite
• LSST: from Science Drivers to Reference Design and Anticipated Data Products
• The Zwicky Transient Facility: System Overview, Performance, and First Results
• The K2 Mission: Characterization and Early results
• The Palomar Transient Factory: System Overview, Performance and First Results
• Kepler Presearch Data Conditioning II - A Bayesian Approach to Systematic Error Correction
• A kilonova as the electromagnetic counterpart to a gravitational-wave source
• Kepler Presearch Data Conditioning I - Architecture and Algorithms for Error Correction in Kepler Light Curves
• The discovery of the electromagnetic counterpart of GW170817: kilonova AT 2017gfo/DLT17ck
• Kepler Flares I. Active and Inactive M dwarfs
• Revised Stellar Properties of Kepler Targets for the Q1-17 (DR25) Transit Detection Run
• Asteroseismology of 16000 Kepler Red Giants: Global Oscillation Parameters, Masses, and Radii
• Automated supervised classification of variable stars I. Methodology
• The Automatic Learning for the Rapid Classification of Events (ALeRCE) Alert Broker
• The Einstein Probe Mission
• RAPID: Early Classification of Explosive Transients using Deep Learning
• Alert Classification for the ALeRCE Broker System: The Light Curve Classifier
• The flare catalog and the flare activity in the Kepler mission
• A data science platform to enable time-domain astronomy
• That's How We Roll: The NASA K2 Mission Science Products and Their Performance Metrics
• Flares in Open Clusters with K2. I. M45 (Pleiades), M44 (Praesepe) and M67
• Deep-Learnt Classification of Light Curves
• Asteroseismology of luminous red giants with Kepler I: Long Period Variables with radial and non-radial modes
• White-Light Flares on Close Binaries Observed with Kepler
• ASTROMER: A transformer-based embedding for the representation of light curves
• Classifying Kepler light curves for 12,000 A and F stars using supervised feature-based machine learning
• TESS Data for Asteroseismology (T'DA) Stellar Variability Classification Pipeline: Set-Up and Application to the Kepler Q9 Data
• Data challenges of time domain astronomy
• Finding flares in Kepler and TESS data with recurrent deep neural networks
• A Bayesian method for detecting stellar flares
• Persistent and occasional: searching for the variable population of the ZTF/4MOST sky using ZTF data release 11
• Accurate and Robust Stellar Rotation Periods catalog for 82771 Kepler stars using deep learning
• Identify Light-Curve Signals with Deep Learning Based Object Detection Algorithm. I. Transit Detection
• Astroconformer: The Prospects of Analyzing Stellar Light Curves with Transformer-Based Deep Learning Models
• Identification of Stellar Flares Using Differential Evolution Template Optimization
• Identifying Light-curve Signals with a Deep Learning Based Object Detection Algorithm. II. A General Light Curve Classification Framework
• The Swan: Data-Driven Inference of Stellar Surface Gravities for Cool Stars from Photometric Light Curves
• A Classifier-Based Approach to Multi-Class Anomaly Detection for Astronomical Transients
• Data challenges as a tool for time-domain astronomy
• Convolutional Neural Networks for Searching Superflares from Pixel-level Data of the Transiting Exoplanet Survey Satellite
• X-ray Sources Classification Using Machine Learning: A Study with EP-WXT Pathfinder LEIA
• Light Curve Classification with DistClassiPy: a new distance-based classifier

Cited by in corpus (1)

• Deep learning-based astronomical multimodal data fusion: A comprehensive review