Name: International Conference on Machine Learning for Astrophysics - ML4Astro
Start: 2022-05-30T08:30:00+02:00
End: 2022-06-01T23:55:00+02:00
Location: Catania

International Conference on Machine Learning for Astrophysics - ML4Astro

from Monday 30 May 2022 (08:30) to Wednesday 1 June 2022 (23:55)

Monday 30 May 2022

08:30

08:30 - 09:15

09:15

09:15 - 09:30
09:30 Session 1
Session 1
09:30 - 10:10
Contributions

09:30 Stellar populations of massive young clusters using supervised machine learning - Koraljka Muzic (CENTRA, University of Lisbon)

09:50 Youth analysis of low-mass stars and brown dwarfs using machine learning methods - Víctor Almendros-Abad (CENTRA, FCUL, Universidade de Lisboa)
10:10 Coffee break
Coffee break
10:10 - 10:40
10:40 Session 2
Session 2
10:40 - 12:10
Contributions

10:40 Stellar dating using chemical clocks and Bayesian inference - Andres Moya (University of Valencia)

11:00 Dust extinction inference from Random Forest regression of insterstellar spectral line widths - Santiago Gonzalez-Gaitan (CENTRA, IST, University of Lisbon)

11:20 QSOs selection in highly unbalanced photometric datasets: The “Michelangelo” reverse-selection method. - Giorgio Calderone (Istituto Nazionale di Astrofisica (INAF))

11:40 Modelling galaxy emission-line kinematics using self-supervised, physics-aware, Bayesian neural networks - James Dawson (Rhodes University)
12:10 Lunch break
Lunch break
12:10 - 14:00
14:00 From Supervised to Unsupervised Machine Learning: lessons learned from learning machines - Kai Polsterer
From Supervised to Unsupervised Machine Learning: lessons learned from learning machines
- Kai Polsterer
14:00 - 14:50
The amount and size of astronomical data-sets was growing rapidly in the last decades. Now, with new technologies and dedicated survey telescopes, the databases are growing even faster. VO-standards provide uniform access to this data. What is still required is a new way to analyze and tools to deal with these large data resources. E.g., common diagnostic diagrams have proven to be good tools to solve questions in the past, but they fail for millions of objects in high dimensional features spaces. Besides dealing with poly-structed and complex data, the time domain has become a new field of scientific interest. By applying technologies from the field of computer sciences, astronomical data can be accessed more efficiently. Machine learning is a key tool to make use of the nowadays freely available datasets. This talk provides an overview of what can be achieved with supervised and unsupervised learning techniques, discussed on examples that show, what we learned when using machine learning algorithms on real astronomical data-set.
14:50 Poster Session
Poster Session
14:50 - 15:20
Contributions

14:50 Planetary Markers in Stellar Spectra: Jupiter-host Star Classification - Miguel Andrea Zammit (University of Malta, Institute of Space Sciences & Astronomy)

14:53 Using Convolutional Neural Networks to Detect and Confirm Exoplanets - Amelia Yu (PAUSD (Palo Alto Unified School District))

14:56 Classification of Evolved Stars with (Unsupervised) Machine Learning - Jamie Welsh (Student)

14:59 Identification of ultracool dwarfs in J-PLUS DR2 using Virtual Observatory tools and Machine Learning techniques - Pedro Mas Buitrago (Centro de Astrobiologia (INTA-CSIC))

15:02 Are large training samples always necessary for machine learning classification models? - Stavros Akras (Institute for Astronomy, Astrophysics, Space Applications and Remote Sensing - National Observatory of Athens)

15:05 Machine learning applied to X-ray spectra: separating stars from active galactic nuclei - Pavan Hebbar (University of Alberta)

15:08 Using Machine Learning to Identify Young Stars with TESS Data - Carlos Santiago (University of California, Irvine)

15:11 Obtaining a classification of A-F stars through clustering analyzing the morphology of the light curves - Jose Ramon Rodon (IAA-CSIC)
15:20 Recent trends in machine learning and opportunities for astrophysics - Concetto Spampinato
Recent trends in machine learning and opportunities for astrophysics
- Concetto Spampinato
15:20 - 16:10
The talk consists of two parts: an overview of the deep learning paradigm from hand-crafted features to learned ones as well as the recent models for classification, regression, object detection semantic segmentation, generation, and their applications including also astrophysics will be given; the second part will instead discuss the recent techniques to understand why deep models make specific predictions diving into the explainable-AI world.
16:10 Session 1
Session 1
16:10 - 16:50
Contributions

16:10 A Convolutional Neural Network to characterise the internal structure of stars - Juan Carlos Suárez (Física Teórica y del Cosmos Dept. University of Granada)

16:30 Deep learning searching for cluster galaxies from multi-band imaging and extensive spectroscopy - Giuseppe Angora (Istituto Nazionale di Astrofisica (INAF))
16:55 Break
Break
16:55 - 17:25
17:25 Session 2
Session 2
17:25 - 18:25
Contributions

17:25 Characterization of Convolutional Neural Networks for the identification of Galaxy-Galaxy Strong Lensing events - Laura Leuzzi (Università di Bologna)

17:45 Unravelling galaxy merger histories with deep learning - Connor Bottrell (Kavli IPMU)

18:05 Multi-band photometry and photo-z estimation from narrow-band images - Laura Cabayol Garcia (IFAE)
Tuesday 31 May 2022

08:30

08:30 - 09:00
09:00 Anomaly Detection in Astronomical Data using Machine Learning - Michelle Lochner
Anomaly Detection in Astronomical Data using Machine Learning
- Michelle Lochner
09:00 - 09:50
The next generation of telescopes such as the SKA and the Vera C. Rubin Observatory will produce enormous data sets, far too large for traditional analysis techniques. Machine learning has proven invaluable in handling large data volumes and automating many tasks traditionally done by human scientists. In this talk, I will discuss how machine learning for anomaly detection can help automate the process of locating unusual astronomical objects in large datasets thus enabling new cosmic discoveries. I will introduce Astronomaly, a general purpose framework for anomaly detection in astronomical data using active learning and overview some recent results.
09:50
09:50 - 11:10
Contributions

09:50 Unsupervised classification reveals new evolutionary pathways - Malgorzata Siudek (IFAE, Barcelona)

10:10 An unsupervised approach to galaxy spectra classification - Julien Dubois (IPAG, UGA, CNRS)

10:30 Patterns in the chaos? An unsupervised view of Galactic SNRs - Cristobal Bordiu (Istituto Nazionale di Astrofisica (INAF))

10:50 Modelling Galactic Microwave emission with ML techniques - Giuseppe Puglisi (Universita' di Roma Tor Vergata)
11:10 Coffee Break
Coffee Break
11:10 - 11:35
11:35
11:35 - 13:00
Contributions

11:35 In Search of the Peculiar: An Unsupervised Approach to Anomaly Detection in the Transient Universe - Dennis Crake (Institute for Astronomy, University of Edinburgh)

11:55 Searching for changing-state AGNs in massive datasets with anomaly detection - Paula Sanchez Saez (ESO)

12:15 Predicting and detecting very-high-energy blazar flares with deep learning - Hermann Stolte (Humboldt University Berlin)

12:35 Galaxy Zoo: Practical Methods for Large-Scale Learning - Mike Walmsley (University of Manchester)
13:00 Lunch Break
Lunch Break
13:00 - 14:30
14:30 Poster Session
Poster Session
14:30 - 15:00
Contributions

14:30 Star- and planet formation caught-in-the-act - Gábor Marton (Konkoly Observatory, Research Centre for Astronomy and Earth Sciences)

14:33 Classification of system variability using a convolutional neural network. - Jozef Magdolen (Advanced Technology Research Institute, Slovak University of Technology in Bratislava)

14:36 Time series anomaly detection in a cataclysmic variable using Support Vector Machine - Denis Benka (Advanced Technologies Research Institute)

14:39 Background Estimation in Fermi Gamma-ray Burst Monitor lightcurves through a Neural Network - Riccardo Crupi (Istituto Nazionale di Astrofisica (INAF))

14:43 Scope: the Zwicky Transient Facility Variable Source Classification Project - Joannes Van roestel (California Institute of Technology/University of Amsterdam) Zwicky Transient Facility collaboration

14:47 Reconstructing blended galaxies with Machine Learning - Lavanya Nemani (Istituto Nazionale di Astrofisica (INAF))

14:50 Estimating Bayesian Posteriors for Galaxy Morphological Parameters using Machine Learning - Aritra Ghosh (Yale University)

14:53 Radio Galaxy detection prediction with ensemble Machine Learning. - Rodrigo Carvajal (Instituto de Astrofísica e Ciências do Espaço, Universidade de Lisboa)

14:57 Radio Image Segmentation with Variational Autoencoders - Hattie Stewart (University of Bristol)
15:00 Time Domain Astroinformatics - Massimo Brescia (Istituto Nazionale di Astrofisica (INAF))
Time Domain Astroinformatics
- Massimo Brescia (Istituto Nazionale di Astrofisica (INAF))
15:00 - 15:50
Astronomy has entered the multi-messenger data era and Data Science has found widespread use in a large variety of applications. The exploitation of synoptic (multi-band and multi-epoch) surveys, like Rubin-LSST, requires an extensive use of automatic methods for data processing and interpretation. With data volumes in the petabyte domain, the discrimination of time-critical information has already exceeded the capabilities of human operators and crowds of scientists have extreme difficulty to manage such amounts of data in multi-dimensional domains. I will introduce the machine/deep learning paradigms suitable to explore Time Domain Astronomy in an efficient and semi-automatic way.
15:50 Session 1
Session 1
15:50 - 16:30
Contributions

15:50 Forecasting solar cycle 25 and next solar minimum using a multi-step Bayesian deep neural network model - Ilaria Bizzarri (Università di Torino)

16:10 Upgrading PlaNET: A Deep Neural Network for Searching for Transiting Exoplanets with the Next Generation Transit Survey - Alexander Chaushev (University of California Irvine)
16:30 Break
Break
16:30 - 17:00
17:00 Session 2
Session 2
17:00 - 18:00
Contributions

17:00 Finding stellar flares with recurrent deep neural networks - Attila Bodi (Konkoly Observatory)

17:20 Classifying Supernova Time-Series Photometry with Convolutional Neural Networks - Helen Qu

17:40 An Unsupervised Dive Into Gamma-ray Burst Afterglow Classification - Eliot Ayache (The Oskar Klein Centre, Department of Astronomy, Stockholm University, AlbaNova, SE-10691 Stockholm, Sweden)
20:00 Dimora De Mauro
Dimora De Mauro
20:00 - 22:00
Wednesday 1 June 2022

08:30

08:30 - 09:00
09:00 Exploring the Universe with the world's largest radio telescope - Grazia Maria Gloria Umana (Istituto Nazionale di Astrofisica (INAF))
Exploring the Universe with the world's largest radio telescope
- Grazia Maria Gloria Umana (Istituto Nazionale di Astrofisica (INAF))
09:00 - 09:30
The Square Kilometre Array (SKA) is an ambitious global science and engineering project aimed at building one of the most impressive astronomical infrastructures ever built. Currently in the construction phase, the SKA will be a network of radio telescopes, distributed over two continents, characterized by high sensitivity and resolving power and by the application of innovative technologies to the development of receivers, to the transport and processing of signal and calculation. Italy, through the National Institute of Astrophysics (INAF), has been one of the major players in the pre-construction phase, participating in the first line in the design, and actively contributes to the definition of scientific cases through the SKA Science Working Groups. Even before the SKA comes online, a series of demonstrator telescopes and systems, known as pathfinders and precursors, are already operational or under development across the world, paving the way for the kinds of technology which the SKA will need to pioneer to make the huge data available to scientists. In this talk information about the project and on its status will be provided. In the framework of the preparatory activity for the SKA, I will also present some scientific highligths from SKA precursors, with particular regards to ASKAP and MeerKAT in the field of Galactic Radioastronomy.
09:30 Machine Learning developments for Radio Astronomy in the SKA era - David Cornu
Machine Learning developments for Radio Astronomy in the SKA era
- David Cornu
09:30 - 10:20
In the context of the preparation of the first SKA observations in a few years from now, more and more public surveys based on precursor and pathfinder instruments are getting released. The analysis of the resulting datasets can already be very challenging as they are getting closer to real upcoming SKA observations. Even if the type of task to perform on such datasets is often rather classical (detection, classification, denoising, etc.), they have become heavily demanding for classical approaches due to datasets size and dimensionality. It is not a surprise then, that many astronomers started to focus their work on Machine Learning approaches that demonstrated their efficiency in similar applications. However, radio-astronomical images are very different from images used to train state-of-the-art pattern recognition algorithms. Moreover, astronomers have specific expectations for the predicted results, be it in terms of robustness, reproducibility, or explainability. As a direct consequence, these methods do not always perform as well as expected when directly applied to astronomical datasets. For this reason, astronomers have started to propose in-depth modifications of widely adopted approaches and also have initiated the development of new dedicated methods. In this talk, I will present an overview of several Machine Learning approaches that have successfully been employed for HI analysis. I will describe methods for a variety of tasks including image denoising, foreground removal, model inversion, etc. but I will put the emphasis on galaxy detection, classification, and characterization techniques, which is a necessary preliminary task for the vast majority of studies. For this, I will present a technical overview of different Machine Learning methods that have been developed by various teams that participated in the SKAO Science Data Challenge 2, which consisted of a 3D detection and characterization task inside a 1TB simulated cube of HI emission. I will discuss the main difficulties identified that are specific to this type of dataset and the various tweaks used by the different teams to mitigate them. I will also discuss what could be future technical developments for these methods in order to overcome the remaining difficulties. Finally, I will present some of the efforts being made in the application of these approaches on pathfinder and precursor instruments, along with the additional difficulties that arise, like the proper definition of learning samples.
10:20 Session 1
Session 1
10:20 - 11:00
Contributions

10:20 Deep Learning Processing and Analysis of Mock Astrophysical Observations - Claudio Gheller (Istituto Nazionale di Astrofisica (INAF))

10:40 Deep neural networks for source detection in radio astronomical maps - Daniel Magro (University of Malta and INAF) Renato Sortino (Istituto Nazionale di Astrofisica (INAF))
11:00 Coffee Break
Coffee Break
11:00 - 11:30
11:30 Session 2
Session 2
11:30 - 12:30
Contributions

11:30 A machine learning classifier for LOFAR radio galaxy cross-matching techniques - Lara Alegre (Institute for Astronomy, University of Edinburgh)

11:50 Analysis of Instance Segmentation Networks for Dispersed FRB Searches - Anastasia Seifert (University of Malta)

12:10 Deep learning 21cm light-cones in 3D - Caroline Heneka (UHH)
12:30 Poster Session
Poster Session
12:30 - 13:00
Contributions

12:30 Effects of incompleteness in the training sample for photoz estimation by DNF algorithm - Laura Toribio San Cipriano (CIEMAT)

12:33 Machine learning investigations for LSST: Strong Lens Mass Modeling and Photometric Redshift estimation - Stefan Schuldt (MPA/TUM)

12:36 Detection of point sources in the CMB intensity using machine learning techniques - Patricia Diego-Palazuelos (Instituto de Física de Cantabria (CSIC-UC))

12:39 Discovering factors that determine dark matter halo abundance with interpretable deep learning - Ningyuan (Lillian) Guo (UCL)

12:43 New applications of Graph Neural Networks in Cosmology - Farida Farsian (Bologna University \ INAF)

12:47 Inferring the Dark Matter halo mass in galaxies from other observables with Machine Learning - Carlo Cannarozzo (Instituto de Astronomía - UNAM)

12:50 Mimicking the halo-galaxy connection using machine learning - Natalí Soler Matubaro de Santi (University of São Paulo)

12:55 Inference of galaxy cluster mass profile using deep learning - Alessia Sbriglio (Università degli studi di Roma "La Sapienza")
13:00 Lunch Break
Lunch Break
13:00 - 14:30
14:30 HPC-cloud convergence is the missing link between scientific computing and applied-AI - Marco Aldinucci
HPC-cloud convergence is the missing link between scientific computing and applied-AI
- Marco Aldinucci
14:30 - 15:20
First, HPC infrastructures are embracing GPUs for their superior performance-per-watt ratio against general-purpose multicores. Second, the next-generation scientific workflows are integrating AI-based steps for their accuracy in approximating and analyzing complex phenomena. Third, AI and specifically Machine Learning (ML), is a perfect workload for GPUs in terms of performance and development time. Today, we cannot still close the circle seamlessly running AI-enabled scientific workloads into HPC infrastructures because their system software and development tools are not designed for modern workloads, such as ML frameworks designed for the cloud. HPC-cloud convergence is likely to bridge the gap. In the talk, we will present Streamflow and CAPIO, two development tools for HPC-cloud convergence.
15:20
15:20 - 16:00
Contributions

15:20 ML Event Reconstructions for Neutrino Telescopes - Philipp Eller

15:40 Gamma/hadron separation for the MAGIC Imaging Atmospheric Cherenkov Telescopes through Convolutional Neural Networks trained with real data samples - Stefano Truzzi (University of Siena/INFN Pisa)
16:00
16:00 - 16:40
Contributions

16:00 GWitchHunters - Machine Learning and Citizen Science to support Gravitational Wave detection - Massimiliano Razzano (University of Pisa and INFN-Pisa)

16:20 Citizen Science and Machine Learning: Towards a Robust Large-Scale Automatic Classification in Astronomy - Manuel Jiménez (Instituto de Astrofísica de Andalucía (IAA-CSIC))
16:40 Break
Break
16:40 - 17:10
17:10 Poster Session
Poster Session
17:10 - 17:40
Contributions

17:10 matryoshka: A suite of neural network based emulators for the power spectrum. - Jamie Donald-McCann (University of Portsmouth)

17:13 Integration and deployment of Model Serving Framework to serve machine learning models at production scale in easy way. - Francesco Caronte

17:16 Predictive Maintenance for a Cherenkov Telescope array - Salvatore Gambadoro (Unict-Inaf)

17:19 Reconstruction and Particle Identification with CYGNO detector - Atul Prajapati (Gran Sasso Science Institute -INFN)

17:23 Deep neural networks for single-line event direction reconstruction in ANTARES - Juan García Méndez (Universidad Politécnica de Valencia)

17:27 Machine-learning based detector optimization of the future P-ONE neutrino telescope - Christian Haack (Technical University of Munich)

17:31 Cats and Dogs vs Gamma and Hadrons - Francesco Visconti (Istituto Nazionale di Astrofisica (INAF))

17:35 Background discrimination for DSNB detection using high-resolution convolutional neural networks - David Maksimović (Johannes Gutenberg-University Mainz)
17:40

17:40 - 18:00