Astro-AI Analytics

Astro-AI Analytics - Research and Education

Astro-AI Analytics is a research group led by Professor Yvonne Becherini (YB) at the Astroparticule et Cosmologie Laboratory (APC) at the Department of Physics at Université Paris Cité.

The group's research activities are dedicated to the application of Artificial Intelligence (AI), Machine Learning (ML), and Data Science in astrophysics, with a focus on multi-messenger astronomy.

In our projects, we combine advanced computational techniques with astrophysical research to optimize data analysis pipelines, extract insights from large datasets, and improve observational strategies for high-energy astrophysics.

YB also teaches PhD-level courses in Applied Data Analytics at Université Paris Cité, besides regular courses in the Physics programme.

Ongoing Projects

CoNIC

The CoNIC project focuses on identifying and classifying astrophysical neutrino sources using machine learning and statistical methods.

By exploiting public and private neutrino event catalogues (e.g., IceCube), CoNIC aims to detect correlations between high-energy neutrinos and active galactic nuclei (AGN).

This approach could provide new insights into the origins of cosmic neutrinos and their connection to extreme astrophysical environments.

ADAPT

ADAPT explores AI-driven approaches to enhance astroparticle data analysis, aiming to reduce reliance on Monte Carlo simulations through Unsupervised and Self-Supervised Learning techniques.

The project focuses on improving event classification, background rejection, and calibration in large-scale experiments such as HESS and KM3NeT.

By integrating these methodologies, ADAPT seeks to streamline data processing workflows, support real-time anomaly detection, and contribute to improving the sensitivity of astrophysical observatories.

SIREX

SIREX explores machine learning and deep learning techniques to improve parameter estimation and data analysis for active galactic nuclei (AGN) and extragalactic sources.

By using simulated data, the project develops AI models for feature extraction and predictive modelling.

SIREX integrates modern machine learning approaches, including Bayesian Neural Networks for time-series analysis, to enhance our ability to infer astrophysical properties from incomplete or complex datasets.

International Collaborations

The HESS (High Energy Stereoscopic System) gamma-ray observatory is a leading ground-based telescope array designed to detect very-high-energy (VHE) gamma rays from astrophysical sources. Located in Namibia, HESS consists of multiple Cherenkov telescopes that capture the brief flashes of Cherenkov light produced when gamma rays interact with Earth's atmosphere. The observatory has made significant contributions to the study of active galactic nuclei (AGN), supernova remnants, pulsar wind nebulae, and the Galactic Centre, helping to map the most energetic processes in the universe. HESS plays a crucial role in multi-messenger astrophysics, complementing observations from space-based telescopes and other cosmic messenger experiments.
KM3NeT (Cubic Kilometre Neutrino Telescope) is an underwater neutrino observatory designed to detect high-energy cosmic neutrinos. Deployed in the deep Mediterranean Sea, KM3NeT consists of vast arrays of optical modules that capture the faint flashes of Cherenkov light produced when neutrinos interact with water. The observatory is structured into two main detectors: ORCA, optimized for studying atmospheric neutrinos and neutrino oscillations, and ARCA, dedicated to identifying astrophysical neutrino sources such as active galactic nuclei (AGN), gamma-ray bursts, and supernovae.
The ALTO CoMET project was a completed research initiative focused on developing a wide-field gamma-ray and cosmic-ray detection array using Cherenkov light detectors. The project, funded at a level of 600k€ by several public and private institutions, including the Crafoord Foundation, included a prototype phase at Linnaeus University in Växjö, Sweden, where detector designs were tested and optimized for sensitivity to very-high-energy gamma rays. ALTO explored innovative methodologies for event reconstruction and background rejection, contributing to advancements in gamma-ray astronomy. The project resulted in two peer-reviewed publications, providing valuable insights that can inform the development of future Cherenkov detector arrays.

Page updated

Google Sites

Report abuse