ARTifical Intelligence For Accelerators,
user Communities and associated Technologies

Unleashing the power of digital transformation in accelerator physics
Dec. 15 2023

Adnan Ghribi\(^{1,2}\)
Adrian Oeftiger\(^{3}\)
Thomas Shea\(^{4}\)

1. GANIL
2. CNRS
3. GSI

Christine Darve\(^{4}\)
Carsten Welsh\(^{5}\)
Jonas L’Haridon\(^{4}\)

4. ESS
5. Cockroft Institute
6. ESF

Sabrina Lecerf\(^{1, 2}\)
Jade Varin\(^{2}\)
Amelia Pollard\(^{7}\)

7. ASTEC STFC
8. Univ. Malta
9. CERN

Gianluca Valentino\(^{8}\)
Lukasz Burdzanowski\(^{9}\)
Andrew Mistry\(^{3}\)

10. DESY
11. INFN

Annika Eichler\(^{10}\)
Verena Kain\(^{9}\)
Stefano Pioli\(^{11}\)

Outline

Introduction

Context, purpose, scope

Now, we need to find the question.

Context

Who’s looking for the question ?

Introduction

Many …

  • Among whom, some are using research infrastructures;

    • Among whom, some are running and building accelerators.

Context

The accelerator landscape : a wide variety

Introduction

More than 30 000 accelerators operational world wide1

More than 99% used in industry and medicine :

  • Industrial applications > 20 000 ;
  • Medical applications > 10 000.

Less than 1% used in research and discovery science :

  • Cyclotrons ;
  • FFA ;
  • Synchrotron light sources ;
  • Linear and circular accelerators/colliders.

Context

The accelerator landscape : a complex picture

Introduction

Purpose

The accelerator landscape : amd many challenges

Introduction

Accelerators pose quasi-industrial challenges in terms of operation and reliability but we need to do things the right way :

  • Improved beam
  • taking into account environmental aspects
  • and transfert to society
  • from the design
  • and through the life time of the machine

Purpose

And how can AI help ?

Introduction

  • operation and reliability ;
  • Detecting, preventing anomalies ;
  • Optimising beam time ;
  • Frugal complex physics simulation ;
  • Improved models.

Several groups have been trying but there are locks to making global impact in the community !

Purpose

What people have been doing

Introduction

Despite all the success stories, there are serious locks that prevent making global impact in the community !

Purpose

so, how do we unlock the use of AI for our RI ?

Introduction

  • We bring the missing piece of FAIRness 
    in data, methods and tools in ML for RI
  • We build upon existing
    knowledge and experience
  • And we push it to its edge(s)
    in an integrated smart pilot/prototype
  • Making sure the challenge stays realistic
    within a given time frame and budget.

Scope

Call

Introduction

  • Name of the call : INFRA-2024-TECH-01-01 ;
  • Call opening : 2023/12/06 ;
  • Call deadline : 2024/03/12 ;
  • Budget/project : 5 to 10 M€
  • Total budget : 63.5 M€

Scope

Aim

Introduction

The aim of this topic is to deliver innovative scientific instrumentation, tools, methods and solutions which advance the state-of-art of RIs in the EU and Associated Countries, and show transformative potential in RIs operation. The related developments, which underpin the provision of improved and advanced services, should lead research infrastructures to support new areas of research and/or a wider community of users, including industrial users. • Cutting-edge technologies will also enhance the potential of RIs to contribute addressing EU policy objectives and socio-economic challenges. • Proposals should ensure complementarity with actions funded under the previous 2022 call (topic HORIZON-INFRA-2022-TECH-01-01 in the 2021-2022 work programme), targeting different instrumentation, tools, methods and solutions. • Proposals should address the following aspects, as relevant: • Research and development of new scientific instrumentation, tools and methods for research infrastructures taking into due account resource efficiency (e.g. energy consumption) and environmental (including climate-related) impacts. This could also include the development of new, more sustainable and efficient methods of collecting data and/or of providing access, including remote and digital, as well as digitalisation of instrumentation, services and results; their technology validation and prototyping training of RI staff for the operation and use of these new solutions. When relevant, developing skills on technical validation to industrial standards; the innovative potential for industrial exploitation of the solutions and/or for the benefits of the society, including facilitating proof of concept for use by SMEs.

Scope

Expected outcomes

Introduction

  • Enhanced scientific competitiveness of Research Infrastructures ;
  • Enhanced RI capacities to address research challenges EU policy priorities ;
  • Increased collaboration of research infrastructures with universities, research organisations and industry ;
  • Increase of technological level of industries through the co-development of advanced technologies of research infrastructures and creation of potential new markets ;
  • Integration of research infrastructures into local, regional and global innovation systems and promotion of entrepreneurial culture.

Scope

Fields of application

Introduction

  • Main field : Accelerator physics and technologies and user communities ;
  • Spans across different applications ;
    • Particle physics ;
    • Nuclear physics ;
    • Light and Neutron sources ;
    • Medical and industrial applications … ;
  • Connects several projects and links to transverse applications.

Consortium

Creation, structuration and crossovers

Some history

Consortium

[March 2023]
First meetings indico

[July 2023]
Workshop at CERN ;
35 participants indico

[Structuration]

[November 2023] Workshop in Paris.
70 participants indico

We gave it a name

Consortium

ARTiFACT

ARTifical Intelligence For Accelerators, user Communities and associated Technologies

This is what came out of it

Consortium

A network, a community

This is what came out of it

Consortium

A network, a community, that extends beyond our walls

  • Industrial partners
  • Observers

Further discussions with potential partners are ongoing.

* Ongoing screening process

This is what came out of it

Consortium

A network, a community, that extends beyond our walls and keeps extending beyond the frame of our projects

Direct connections and transverse contributions to several projects.

This is what came out of it

A pool of study cases

Consortium

  • Over 50 study cases ;
  • With a wide variety of applications ;
  • Some emerging as driving cases ;
  • Other gravitating to ensure a wide impact in the community.

This is what came out of it

A pool of study cases

Consortium

Project

Strategy, organisation, planning

Strategy

Project

From chaos, a pattern emerges

  • Standardisation
  • Best practices
  • Compliance
  • Documentation
  • Interoperability
  • Environment considerations
  • Ethical considerations

Strategy

Project

From chaos, a pattern emerges

  • FAIR1ness in data, methods, tools, models, …
    • from production (raw data) ;
    • to curation ;
    • to structuration ;
    • to publication ;
    • to model training ;
    • to deployment.

Strategy

Project

From chaos, a pattern emerges

  • A bridge between :
    • New methods and innovative tools ;
    • Operation of exiting RI and design of the future ones ;
    • Trustworthy, ethical and community driven developments ;
    • Challenging projects through the use-cases organisation.

Strategy

Project

From chaos, a pattern emerges

  • Real life integration in test facilities ;
  • Partial integration and validation foreseen at different facilities ;
  • Two test facilities identified for full integration :
    • TEX & CLARA.

Strategy

Project

From chaos, a pattern emerges

  • Although enabling actions exist, all developments will be progressing concurrently ;
  • Transverse actions like knowledge transfer and integrating activities will also link the different work-packages.

Governing structure

Project

From the pattern, clusters aggregate

WP names are subject to change according to the ongoing discussions.

WP.1 Management

Project organisation

GANIL | CNRS | GSI | ESS

Contact for more
information and
contributions

Administrative contact :
Sabrina Lecerf
EU office contact :
Jade Varin
Scientific coordination :
Adnan Ghribi
Adrian Oeftiger
Thomas Shea

  • Scientific coordinator + 2 deputies
    • GANIL/CNRS - GSI - ESS
  • Project coordination and office
    • GANIL
  • Integrates an innovation fund.

WP.2 Integration Activities

Project organisation

Communication, Dissemination and Industry Engagement

Jonas L’Haridon (ESF) / Emmanuel Detsi (ESF)

Objectives

  • Disseminate and communicate the project outputs ;
  • Set-up and maintain multiple communication channels ;
  • Consolidate project information results ;
  • Maximize the impact of the project.

Participants

ESF ; ESRF ; ESS ; Delft Univ.

Key Tasks

  • Dissemination, exploitation and communication ;
  • Stakeholders mapping and industry engagement ;
  • Integration within relevant EC roadmaps ;
  • Public engagement events ;
  • Roadmap on exploitation opportunities.

WP.3 Training and knowledge transfer

Project organisation

Carsten Welsh (Cockroft Institute) / Christine Darve (ESS)

Motivation

Develop activities and platforms to train experts and support knowledge transfer.

Participants

Cockroft Institute, ESS, GSI, Hopswork

Objectives

  • Track progress across other WPs to implement relevant training content and syllabus ;
  • Prepare materials (e.g. MOOC, background, videos) and disseminate in schools, workshops and institutes webpages ;
  • Prepare exchange programs related to other workpackages tasks (professional training and expertise exchange) ;
  • Prepare and implement Hackathons and data challenges ;
  • Incorporate ethics and responsible data handling (UN 17 Sustainable development goals).

WP.4 AI Guidelines for accelerators

Project organisation

Gianluca Valentino (Univ. Malta)

Motivation

Provided an integrated guideline for AI enabled/ready design, control, operation and data handling.

Participants

University of Malta ; CERN ; GSI ; Hopswork (ind) ; University of Granada ; Cosylab (ind) ; IFMIF-DONES

Tasks and deliverables

  1. Identify AI boundaries and requirements for RIs ;
    Readiness assessment and requirements analysis.
  2. Define guidelines for AI-ready accelerator data infrastructure ;
    Guidelines for AI-ready data infrastructure.
  3. Define guidelines for AI-ready controls infrastructure ; Guidelines for AI-ready controls infrastructure.
  4. Define guidelines for training of RI staff to enable AI adoption ;
    Guidelines for training
    Adoption assessment

WP.5 AI ready accelerator data on the cloud

Project organisation

Andrew K. Mistry (GSI) / Jim Downling (Hopswork)

WP.5 AI ready accelerator data on the cloud

Project organisation

Andrew K. Mistry (GSI) / Jim Downling (Hopswork)

Motivation

  • Develop a data curation and stucturation tool to homogenise data making it FAIR and ML ready
  • Establish a prototype platform that enables data access and processing

Participation

GSI/FAIR, Hopswork, IJClab, Soleil, ESS, DMSC, Cosylab, PaNOSC, ExPANDS, ESCAPE, (CERN), (EOSC), CC IN2P3, Centrale Supelec

Tasks and deliverables

  1. Development of a Data generation and Curation Generalised framework.
  2. Development of an Agnostic Tool for Data Curation, Structuration, and Exploration.
  3. Testing tool and deployment.
  4. Openness and Community development of the tool.
    DMP ; Data catalog, Agnostic curation/homogenisation tool, ML enabled opensource platform

WP.6 AI algorithms in the accelerator context

Project organisation

Annika Eichler (DESY) ; Verena Kain (CERN)

Motivation

Develop algorithms, tools and methods and interface them with simulation tools and command/control.

Participation

DESY, CERN, KIT, GSI, LPHC, Univ. Liverpool, IJCLab, IFMIF-DONES, ESS, ESS-Bilbao, GANIL, EPFL, ESRF, Soleil.

Tasks status (5 tasks)

  1. Methods/requirements exploration for accelerator sciences ;
  2. Software framework development ;
  3. Control and optimization ;
  4. Anomaly detection, prediction and prevention ;
  5. Modelling and future RI.

WP.6 AI algorithms in the accelerator context

Project organisation

Annika Eichler (DESY) ; Verena Kain (CERN)

WP.7 Agile smart accelerator prototype

Project organisation

Stefano Pioli (INFN) ; Amelia Pollard (ASTEC STFC)

Motivation

  • Demonstrate tools and techniques from other work packages.
  • TEX and CLARA for system level integration demonstration.
  • Distributed sub-system level integration demonstration.

Participation

INFN(Frascati), ASTEC STFC, Soleil, ESRF

Tasks

  1. Develop infrastucture for ML model training and deployment ;
  2. Develop infrastructure for data collection at high repetition rates and for long term storage and sharing ;
  3. Develop API for high-level machine parameter control ;
  4. Develop generalised system for passing monitoring data to models for online optimisation ;
  5. Track progress across other work packages and provide technical input.

Risk mitigation

Project organisation

  • No unique dependency on one partner, lab or company1 ;
  • Leaders and co-leaders have complimentary roles ;
  • Balance between participants (gender, age, country) ;
  • Inputs and outputs present different levels of TRL2 ;
  • Administrative support ready from the start with a EU project team ;
  • External Advisory board of observers planned.

Next steps

Project organisation

  • Excellence
    • Clarity and pertinence of the project’s objectives and methodology, and the extent to which the proposed work is ambitious, and goes beyond the state of the art.
  • Impact
    • Credibility of the pathways to achieve the expected outcomes and impacts ;
    • Suitability and quality of the measures to maximise expected outcomes and impacts (ex: dissemination and exploitation plan).
  • Implementation
    • Quality and effectiveness of the work plan, assessment of risks, and appropriateness of the effort assigned to work packages, and the resources overall.

Conclusion

  • A lot has been done in very little time thanks to the proactive involvement of all participants ;
  • Preparation of the project submission is progressing as planned ;
  • A very nice energy in the different working groups

This is not a black box magic, this is math.

This is not utopia, this is a community driven dev, and this is happening.

Thank you

Questions ?

References

Press F for full screen | M for menu | ? for help | \(\rightarrow\) to change slide

Chou, Weiren, and Alexander Wu Chao. 2019. Reviews of Accelerator Science and Technology-Volume 10: The Future of Accelerators. World Scientific.