Physical Simulations — Model HR
H-rich semi-analytical explosions; multi-band light curves with cadence and uncertainty metadata.
Last updated: 31/08/2025
ESPAI
Advanced Supernova Transient Research with Artificial Intelligence
ASTRAI advances the understanding of Supernovae by combining innovative astrophysical models with Artificial Intelligence to generate synthetic datasets, enable automated data characterisation, and provide open, scalable tools for the era of Big Data in astronomy.
ASTRAI has been selected and funded under the ICSC – Centro Nazionale di Ricerca in HPC, Big Data e Quantum Computing, as part of the PNRR, and supported by the European Union – NextGenerationEU.
Access Resources
Learn More
Welcome to the ASTRAI Website
This website introduces the ASTRAI project — what it is, what it does, and why it matters. You’ll also find access to the project’s main outputs, from datasets and AI models to source code and publications. Explore our resources to dive into cutting-edge tools for supernova characterisation and synthetic data generation.
Everything is open and accessible. See the Outputs section below to learn more, or open the ASTRAI Resource Hub to access the project resources. If you’d like to download datasets or code, we’ll just ask you to fill out a short form with your affiliation and intended use.
Find collaborations and related work with references and external links.
Discover the Project
Learn in detail about ASTRAI, its scientific goals, and the motivation behind using Artificial Intelligence to study supernovae. Understand how the project combines astrophysical models with advanced machine learning techniques to create new opportunities for research and discovery.
Learn MoreAccess Resources
Browse a growing collection of datasets, source code, and publications as they are released. These resources are designed to support both the astrophysics and AI communities, offering practical tools for analysis, experimentation, and reproducible science.
ExploreCheck Usage Terms
Find clear and transparent information about licensing conditions, recommended citation formats, and guidelines for the responsible reuse of materials. These terms help ensure that ASTRAI’s outputs are properly credited and can be applied ethically in future research.
View TermsAbout the ASTRAI Project
ASTRAI is a research initiative dedicated to the automatic characterisation of supernovae, combining physics‑based modelling with modern machine learning. Supernovae shape the chemical enrichment of galaxies, influence stellar evolution, and serve as distance indicators, yet their study is challenged by complex explosion mechanisms and irregular observations.
To address these challenges, ASTRAI integrates semi-analytical models of supernova physics with advanced AI methods, creating synthetic datasets that reproduce realistic light curves, temperatures, and velocities. These resources enable the training and validation of machine learning models capable of extracting physical parameters from both real and simulated data, even under observational uncertainty.
The project delivers openly accessible tools: curated datasets, generative frameworks, parameter-recovery algorithms, and source code repositories, all provided under transparent licensing terms. In this way, ASTRAI contributes to reproducible science and accelerates discovery for both the astrophysics and AI communities.
🧠 AI for Transients
Models tailored to supernova light curves and spectra that infer meaningful physical parameters from sparse, noisy and unevenly sampled data.
🔬 Physics‑Informed Modelling
Semi‑analytical descriptions of explosion and ejecta evolution that help constrain temperatures, velocities and luminosities across different supernova classes.
🧪 High‑Fidelity Synthetic Data
Generative pipelines that create realistic light curves and related observables, bridging gaps in coverage and strengthening training and evaluation.
Project outputs
- Physical models: Documented semi-analytic descriptions for H-rich and interacting supernovae. The models specify the main physical parameters (ejecta mass, kinetic energy, progenitor radius, radioactive contribution, circumstellar-medium properties) and produce time-dependent observables including multi-band light curves, temperature and photospheric velocity, together with stated assumptions and parameter ranges.
- Synthetic datasets: High-fidelity light-curve sets generated from controlled physical parameters and observational effects. Each sample includes metadata (passbands, cadence, uncertainty model, phase coverage, redshift if applicable), links to the generating parameters, train/validation/test splits, and quality flags to support benchmarking and method comparison.
- Inference and characterisation tools: Estimators that recover physically meaningful parameters from real and synthetic observations with quantified uncertainty. They are robust to irregular sampling and noise, and are accompanied by calibration checks (e.g. coverage of credible intervals), ablation/robustness tests, and reference baselines for comparison.
Validation & benchmarking
Consistency checks against semi-analytic expectations, posterior-predictive tests on synthetic samples, and comparisons with observed light curves where available. Performance reporting focuses on parameter-recovery accuracy, timing features (e.g. rise/plateau epochs), and uncertainty calibration, enabling transparent, reproducible evaluation across methods and datasets.
Methodology at a glance
A unified pipeline links (i) physics-informed modelling, (ii) controlled synthetic data generation, and (iii) uncertainty-aware AI characterisation. The workflow supports multi-band time series with irregular cadence and scales to large experiments on HPC resources, so models can be trained, stress-tested, and interpreted in a consistent manner.
Ownership, collaboration & funding
ASTRAI is conceived and developed under the leadership of Koexai S.r.l., which coordinates all project activities and ensures their delivery. The project benefits from the valuable scientific input of Prof. Maria Letizia Pumo and her team at the University of Catania, who contribute expertise in physical modelling and act as scientific referees, nominated by INAF (Spoke 3 leader).
ASTRAI has been selected and funded under the ICSC – Centro Nazionale di Ricerca in HPC, Big Data e Quantum Computing, as part of the PNRR, and supported by the European Union – NextGenerationEU.
Outputs
Explore ASTRAI’s outputs: datasets, source code, publications, diagrams, and findings. For downloads and full details, open the Resource Hub.
Physical Simulations — Model INT
Interacting CSM scenarios; synthetic light curves with parameter provenance.
Last updated: 31/08/2025
Real Supernova Observations
Curated multi-band observations with passbands, cadence, uncertainty model and quality flags.
Last updated: 31/08/2025
GenAI Simulator Dataset
Synthetic light curves generated by the GenAI simulator pipeline.
Sample preview
First 20 rows. Columns and units match the full dataset.
Loading preview…
Download sample CSV
Checksum (SHA-256):
TBC
Last updated: 31/08/2025
Requesting the full datasets
Large files & licensing: full releases are provided for non-commercial research on a short request.
- Tell us your name, affiliation, email, and intended use.
- We’ll reply with access instructions and a download link.
Contact & Acknowledgements
Project Contact
For resource access requests, collaboration inquiries, or technical questions:
info@koexai.comOwnership & Scientific Referees
Koexai S.r.l. — Project lead and coordinating body.
University of Catania (Prof. Maria Letizia Pumo) — Scientific referee and contributor, nominated by INAF (Spoke 3 leader).