Scientific Calendar Event



Description
An ICTP Meeting 'in person'

Machine learning (ML) is becoming an increasingly essential tool to advance research in fields as diverse as Astronomy, Biophysics, Particle Physics, Environmental science, and Material Science. By extracting patterns and information from large and complex data sets, ML helps in understanding complex multi-scale phenomena that would otherwise be intractable.  The development of efficient, robust, interpretable and reproducible ML pipelines is thus crucial to advance scientific discovery in computational sciences.
 
We are excited to announce an the Advanced School in Applied ML at ICTP,  which will take place in Trieste (Italy) from May 27th to June 1st, 2024, dedicated to PhD students and junior researchers. The school aims to bridge the gap between theoretical notions in data science/ML and application of ML to cutting edge scientific problems, with a focus on scalability and advanced HPC methods for ML, and in particular for deep learning (DL). 
 
Lecturers from top-ranked research centers/universities and big tech companies (e.g. Intel, Nvidia, IBM) will cover the following topics: symbolic regression, Gaussian processes for time-series data analysis, transfer learning and domain adaptation, Graph Neural Networks and Transformers for multi-modal analysis, AI explainability, Neuromorphic computing, for energy-efficient DL training. Each topic will be complemented by the necessary tools to create efficient, scalable, and portable ML pipelines that exploit the capabilities of modern HPC infrastructures.
 
30 participants will be selected with attention to creating a cohort with balanced representation of countries (particularly, least developed countries), gender and under-represented groups.  
 
This school will be the first activity in the field of advanced applied Machine Learning organised by the brand-new ICTP International Consortium for Scientific Computing (ICSC).

Speakers:
Matteo Angelinelli, CINECA
Cristiano De Nobili, PI School
Iacolo Colonnelli, University of Turin
​Gabriel Fonseca-Guerra, INTEL
Jarvist Frost, Imperial College London
Caroline Heneka, University of Heidelberg
Gianluca Mittoner, University of Turin
Emanuele Panizon, Area Science Park
Nabeel Seedat, Cambridge
​Savannah Thais, Columbia University
Paolo Viviani, LINKS foundation/B-CRATOS


Poster abstract: Applicants are encouraged to submit an abstract for a poster session. Please use the ICTP templates that are available below for download.

We would also suggest to apply to the following activity that may also be of interest to you: "Youth in High Dimensions: Recent Progress in Machine Learning, High-Dimensional Statistics and Inference".

Grants: A limited number of grants are available to support the attendance of selected participants, with priority given to participants from developing countries. There is no registration fee.
 

Go to day
  • Monday, 27 May 2024
    • 08:00 - 09:00 Registration
      The registration is necessary for participants who does not stay at the ICTP Guest Houses.
      Location: Lobby at the Entrance (Leonardo building)
    • 09:00 - 17:00 The Future of Scientific Computing: A Global Perspective
      Please consult the Symposium Overview web page HERE and Symposium Programme HERE
      
      The attendance is required.
      Location: Budinich Lecture Hall (Leonardo Building)
  • Tuesday, 28 May 2024
    • 08:30 - 17:30
      • 08:30 Opening 30'
        Speaker: Serafina DI GIOIA (ICTP), Ivan GIROTTO (ICTP), Roberto TROTTA (Imperial College London, SISSA)
      • 09:00 ML for Astrophysics 1h30'
        Speaker: Caroline HENEKA (University of Heidelberg, Germany)
      • 10:30 Coffee Break 30'
      • 11:00 ML for material science 1h30'
        Speaker: Jarvist FROST (Imperial College London, UK)
        Material: Abstract
      • 12:30 Lunch 1h30'
      • 14:00 Tutorial on "ML analysis of astronomical data" 1h30'
        Speaker: Caroline HENEKA (University of Heidelberg, Germany)
      • 15:30 Coffee Break 15'
      • 16:00 Tutorial on "Diffusion models and GP-based ML" 1h30'
        Speaker: Jarvist FROST (Imperial College London, UK)
  • Wednesday, 29 May 2024
    • 10:00 - 17:30
      • 10:00 Coffee Break 15'
      • 10:15 Data-centric ML 1h30'
        Speaker: Nabeel SEEDAT (Cambridge, UK)
      • 11:45 Group photo 15'
      • 12:00 Lunch 1h0'
      • 13:15 Dynamic hybrid workflow for DL on HPC infrastructure 1h30'
        Speaker: Iacopo COLONNELLI (University of Turin, Italy)
      • 14:45 Coffee break 30'
      • 15:15 Intro to Federated learning 1h30'
        Speaker: Gianluca MITTONE (University of Turin, Italy)
    • 19:00 - 20:30 Social dinner
      Location: Trieste ( City centre )
  • Thursday, 30 May 2024
    • 09:00 - 17:30
      • 09:00 Intro to GNNs 1h30'
        Speaker: Savannah J. THAIS (Columbia University, USA)
      • 10:30 Coffee Break 30'
      • 11:00 Tutorial on GNNs 1h30'
        Speaker: Savannah J. THAIS (Columbia University, USA)
      • 12:30 Lunch 1h30'
      • 14:00 AI applied to meteorological research 1h30'
        Speaker: Matteo ANGELINELLI (CINECA)
        Material: Abstract
      • 15:30 Coffee Break 15'
      • 16:00 Poster session 1h30'
        Material: Booklet
  • Friday, 31 May 2024
    • 09:00 - 17:15
      • 09:00 Online learning for smart prosthetics: B-CRATOS 30'
        Online
        Speaker: Paolo VIVIANI (LINKS foundation/B-CRATOS, Italy)
        Material: Description
      • 09:30 RL - from zero to hero 1h0'
        Speaker: Emanuele PANIZON (Area Science Park, Italy)
        Material: Description
      • 10:30 Coffee Break 30'
      • 11:00 RL - from zero to hero 1h30'
        Speaker: Emanuele PANIZON (Area Science Park, Italy)
      • 12:30 Lunch 1h30'
      • 14:00 Intro to Neuromorphic computing 1h30'
        Speaker: Gabriel Andres FONSECA GUERRA (Intel Corporation, Germany)
      • 15:30 Coffee Break 15'
      • 15:45 Tutorial: Intro to LAVA / training a DL model on Loihi 2 1h30'
        Speaker: Gabriel Andres FONSECA GUERRA (Intel Corporation, Germany)
  • Saturday, 1 June 2024
    • 09:00 - 19:00
      • 09:00 From zero to the Transformer Architecture in NLP 1h30'
        Speaker: Cristiano DE NOBILI (University of Milan Bicocca, Italy)
        Material: Description
      • 10:30 Coffee Break 30'
      • 11:00 From the Transformer to GPT-2 and Small Language Models (Part 1) 1h30'
      • 12:30 Lunch 1h30'
      • 14:00 From the Transformer to GPT-2 and Small Language Models (Part 2) 1h30'
      • 15:30 Coffee Break 15'
      • 16:00 Playing with open-source LLMs 1h30'
      • 17:30 Parallel ML (Dask vs Spark for Parallel I/O) 1h30'