Code Repository
A complete, documented repository containing all components of the GRAIS pipeline — preprocessing tools, simulation scripts, deep-learning models, and validation utilities.
Last updated: 30/11/2025
GRAIS
Gamma Ray Artificial Intelligence System.
GRAIS advances the detection and simulation of Gamma Ray Bursts (GRBs) by combining innovative astrophysical models with Artificial Intelligence to generate synthetic datasets, enable automated data detection, and provide open, scalable tools for the era of Big Data in astronomy.
GRAIS 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 GRAIS Website
This website introduces the GRAIS 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 GRBs detection and synthetic data generation.
Everything is open and accessible. See the Outputs section below to learn more, or open the GRAIS 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.
Discover the Project
Learn in detail about GRAIS, its scientific goals, and the motivation behind using Artificial Intelligence to study GRBs. 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 GRAIS's outputs are properly credited and can be applied ethically in future research.
View TermsAbout the GRAIS Project
GRAIS is a research initiative dedicated to the automatic detection of GRBs, combining physics‑based modelling with modern machine learning. GRBs are really rare and unique events, and the detection is not always easy to achieve.
To address these challenges, GRAIS integrates deep learning models with state-of-the-art astrophysical frameworks like fermitools, creating synthetic datasets that reproduce realistic GRBs, times and energies. These resources enable the training and validation of machine learning models capable of extracting physical features from both real and simulated data.
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, GRAIS contributes to reproducible science and accelerates discovery for both the astrophysics and AI communities.
🧠 AI for Transients
Machine-learning models designed to detect rare and faint astrophysical events hidden in noisy gamma-ray data.
🔬 Physics‑Informed Modelling
AI methods guided by physical constraints and LAT instrument response, ensuring reliable and interpretable detections.
🧪 High‑Fidelity Synthetic Data
Realistic simulations of GRB afterglows and background conditions to train, validate, and stress-test the pipeline.
Project outputs
- Project outputs: GRAIS delivers a fully automated pipeline for discovering GRBs in the Fermi LAT archive. The project produces advanced deep-learning models, physics-informed detection tools and a curated collection of labelled candidate events. These outputs enable researchers to explore the high-energy sky with unprecedented sensitivity, improving our ability to identify rare signals that are typically missed by standard analysis techniques.
- Synthetic datasets: To train and validate the AI models, GRAIS generates high-fidelity synthetic datasets that reproduce the complexity of real LAT observations. These datasets combine physically accurate GRB simulations, realistic background modelling and controlled astrophysical scenarios. They allow systematic testing of the pipeline, the creation of large balanced training sets, and the exploration of edge cases that rarely occur in the real sky.
- Inference: The inference engine applies the trained models to the full LAT photon data stream, scanning billions of events with high speed and precision. It evaluates spatial, temporal and spectral patterns to identify anomalous photon clusters compatible with transient sources. The system is optimised for operational reliability and can be used to support rapid discovery, catalogue expansion, and follow-up analysis.
Validation & benchmarking
GRAIS is tested using both real LAT data and controlled synthetic injections to assess sensitivity, false positives, and localisation accuracy. Performance is benchmarked against standard methods, ensuring a reliable and reproducible detection pipeline.
Methodology at a glance
GRAIS integrates modern AI techniques with domain-driven insights to search for elusive gamma-ray transients. The system analyses large volumes of observational data, highlights unusual patterns, and filters them through dedicated models to separate real astrophysical events from noise. This streamlined workflow enables fast, reliable discovery across the full mission dataset.
Ownership, collaboration & funding
GRAIS 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 support of Francesco Longo and Sara Cutini who acted as scientific referees, nominated by INAF (Spoke 3 leader), and from the scientific input of Riccardo Martinelli who has shared his expertise in the astrophysical context too.
GRAIS 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 GRAIS outputs: datasets, source code, publications, diagrams, and findings.For downloads and full details, open the Resource Hub.
Synthetic Dataset
A high-fidelity collection of simulated GRBs, generated using FermiTools–based physical simulations and complementary GenAI models.
Last updated: 30/11/2025
GRB Candidate Catalogue
A curated set of transient candidates identified by the anomaly-detection pipeline.
Last updated: 30/11/2025
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, Collaborations & Funding
Koexai S.r.l. — Project lead and coordinating body.
Francesco Longo — Scientific referee, nominated by INAF (Spoke 3 leader).
Sara Cutini — Scientific referee, nominated by INAF (Spoke 3 leader).
Riccardo Martinelli —Ph.D student in university of Trieste