4th Community Workshop 2024 &
AIDA Symposium
26 – 27 June 2024
Venue: Centre for Research and Technology Hellas – CERTH
Location: Thessaloniki, Greece
To celebrate four years of building excellence of AI research and foresighting the impact of the European Networks of Excellence Centres in AI and Robotics (AI NoEs), the Fourth Community Workshop 2024 aims to provide a forum for sharing ideas and best practices and sustaining collaborative activities beyond the NoEs’ lifetime.
The event is open for a broader audience beyond the NoEs community – simply register for online attendance and join the exchanges with us!
The two-day 4th Community Workshop is hosted by CERTH and co-organised by all the NoEs, with one specific day on AI and education.
● 26 June 2024:
NoE Community Workshop
● 27 June 2024:
AIDA Symposium “AI Education Beyond Borders”
The goal is to foster collaborations within the NoEs’ community and to discuss ways forward to maximise the impact of NoE research in 2025 – and beyond. This 4th Community Workshop builds on the previous events:
the First ICT-48 Community Workshop held online in June 2021, a very fruitful Second ICT-48 Community Workshop held in Brussels in October 2022, and the Third Community Workshop in June 2023 Siena.
Target Audience
Online Attendance:
The workshop is designed specifically for the entire AI NoEs’ Community but is open to anyone interested. In this case, please register for an online participation only. This registration form can also be used by members of the NoEs community who cannot (for whatever reason) attend onsite.
For remote listeners at the plenary sessions, the possibility of participation is unlimited across the broader community and beyond.
Onsite Attendance:
Due to limited room capacity, the workshop is open to a given number of NoEs community members to attend on-site. To ensure a balanced group of in-person participants from NoEs stakeholders, each NoE will receive 8 tickets. A link to a special registration form will be emailed separately to NoEs coordinators and project managers so that they can decide on on-site participation for each NoE and distribute the link.
Also the representatives of the European Commision, ADRA, ADRA-e and AI4Europe are more then welcome and expected to come to Thessaloniki.
Day 1: Agenda
NoE Community Workshop
This is the draft agenda. The topics of some sessions will be specified soon as well as the list of parallel workshops (relevant for onsite participation mainly).
Wed | 26 June 2024
NoE Community Workshop
8:30 – 9:00 Walk-in
9:00 – 9:10 Welcome
Chair: Yiannis Kompatsiaris (AI4Media)
9:10 – 9:45 Keynote on the future of AI research & innovation
Lucilla Sioli, EC DG CONNECT (online)
9:45 – 11:00 Highlights of the NoEs and introduction of the new NoEs (dAIEDGE, ELIAS, ENFIELD)
Chair: Nicu Sebe
Speakers: 1 representative per NoE
11:00 – 11:30 Coffee Break
11:30-13:00 Parallel Workshops 1-3
Parallel Workshop 1: Evaluating and Auditing, versus Safe and Secure by Design AI Systems
Organisers: ELSA and ENFIELD
Parallel Workshop 2: Navigating Interdisciplinarity
Organisers: AI4Media
Parallel Workshop 3: #EuroGen: Mapping the Future with Generative AI
Organisers: euROBIN and HumanE-AI-NET
13:00 – 14:00 Lunch Break
14:00 – 15:30 Parallel Workshops 4-6:
Parallel Workshop 4: Next-generation infrastructure federation: a Virtual research lab for edge AI
Organiser: dAIEDGE
Parallel Workshop 5: Assessing and Enhancing Fairness in AI Systems
Organisers: AI4Media and ELIAS
Parallel Workshop 6: NoE collaborative actions: Joint SRA and Ecosystem Mapping
Organisers: VISION, TAILOR and ELISE
15:30 – 16:00 Coffee Break
16:00 – 17:00 Worldcafé on sustainability of the NoE community
17:00 – 17:30 Closing town hall
20:30 – 22:00 Dinner
Day 2: Agenda
AIDA Symposium: AI Education Beyond Borders
This is the draft agenda. The topics of some sessions will be specified soon.
Thu | 27 June 2024
8:30 – 8:50 Walk-in
8:50 – 9:00 Welcome
Chair: Ioannis Pitas (AI4Media)
9:00 – 10:00 AI University Education
Organiser: Efstratios Gavves, University of Amsterdam
Panelists: Evangelos Kanoulas, University of Amsterdam; Vincenzo Lomonaco, University of Pisa
In this panel, we will discuss how AI already impact and will continue impacting higher education, not only in AI and computational sciences but also natural sciences and humanities. What are the building blocks of AI, which of them will be critically needed for other sciences, what is the best way to deliver this education? In this panel we will be focusing on all these questions.
10:00 – 11:00 Training AI Expertise for Europe: Initiatives and the EU AI PhD Curriculum
Organiser: Barry O’Sullivan, University College Cork | VISION
Panelists: Ioannis Pitas, Aristotle University of Thessaloniki | AI4Media; Fredrik Heinz, Linköping University | TAILOR; Peter Flach, University of Bristol (remote); Mohamed Chetouani, ISIR (remote).
Doctoral-level Artificial Intelligence (AI) education is pivotal in shaping the next generation of AI professionals across various domains. The session focuses on the efforts within the European AI community to train AI expertise for the EU and a joint effort to establish a standardized and comprehensive AI PhD curriculum.
Spearheaded by the European networks of excellence (NOEs or ICT-48 projects) namely AI4Media, ELISE, HumanE-AI Net, TAILOR and VISION.
11:00 – 11:30 Coffee Break
11:30-12:30 EBU Academy School of AI: How EBU Academy is upskilling Europe’s Public Service Media on AI
Organiser: Justin Kings, EBU Academy (remote)
Panelists: Hélène Rauby Matta, EBU Academy (remote)
Artificial Intelligence will bring profound change for Public Service Media across Europe and the rest of the world. AI offers exciting opportunities in areas such as content creation and journalism but it also brings new ethical and organizational challenges. In this session, you can learn how the EBU Academy, the learning and development unit of the European Broadcasting Union, is responding to the media industry’s need for acquiring critical knowledge and training, by opening its own School of AI.
12:30-13:00 AIDA and AIoD platform and experience
Organiser: Long Pham, University College Cork | VISION
Panelists: Stefano Berretti, University of Firenze
13:00 – 14:00 Lunch Break
14:00 – 15:00 Track 1: Innovation for Education: How AI can revamp long-life learning
Organiser: Salvatore Moccia, Head of Education and Skills at EIT Digital
Panelists: Vanessa Wainwright, Abodoo; Nektar Baziotis, ICARUS AI E-Learning
Since ChatGPT was released late 2022, followed swiftly by other “large language models”, hype and concern about AI’s possible impact on education has flooded the media. In response to “generative AI” applications arriving in Universities and Schools, the Assistant Director-General of Education at UNESCO, Stefania Giannini, wrote that “The speed at which generative AI technologies are being integrated into education systems in the absence of checks, rules or regulations, is astonishing”. (Education in the Age of Artificial Intelligence, Octobre 2023).
During this panel, we would like to explore the other uses of AI in education, apart from generating texts. What are the possible applications of AI in Education? What impact they can have? What the future of education will look like? How upskilling and reskilling can be impacted by AI?
14:00 – 15:00 Track 2: AI in journalism and misinformation: Overview, audiovisual applications and media literacy needs
Organiser: Nikos Sarris CERTH | AI4Media
Panelists: Andreas Veglis, Charalampos, Dimoulas and Nikolaos Vryzas, Aristotle University of Thessaloniki
This panel will discuss the latest trends in AI and particularly generative AI technologies that (may) affect journalism in positive and negative ways. Particular tools that have been developed in this area and can be helpful to journalists, fact-checkers and researchers, will be presented and demonstrated, discussing the challenges and opportunities for future developments.
15:00 – 15:30 Coffee Break
15:30 – 16:30 AI tools in education (impact of LLMs and Generative AI in education, exams, essays)
Organiser: Efstratios Gavves, Marc Salomon, University of Amsterdam
Panelists: George Tsatsaronis, Elsevier; Kristen DiCerbo, Khan Academy
In this panel, we will focus on the impact of AI tools on education, its positive uses that can make education more effective and personalized, as well as the possible negative impact on examination and quality assurance.
16:30 – 17:30 ICT-48 Sustainability Session with the EC representatives
17:30 – 18:30 AIDA General Assembly
Registration
For those who will attend ONLINE only
With the exception of the parallel sessions on the afternoon of the first day, most of the event programme will be live streamed and open to all those interested in the future of European AI and robotics. Likewise, all other partners and members of the NoEs who are not attending in person are welcome and encouraged to participate remotely.
Feel free to register for either or both days and we will send you a link for your online participation shortly before the community workshop.
Venue:
Please join us in the Conference Centre of Centre for Research and Technology Hellas (CERTH).
CERTH (“EKETA” in Greek)
6th km Charilaou – Thermi Road
570 01 Thermi
Thessaloniki, GREECE
CERTH is located at the 6th km of the Charilaou-Thermi Road in Thermi, Thessaloniki. It is about 10km from the centre of Thessaloniki and about 8 km from “Makedonia” International airport.
More about the Agenda
Thursday 26 June 2024: NoEs Community Workshop – Parallel Workshops
11:30 – 13:30
PW 1: Evaluating and Auditing, versus Safe and Secure by Design AI Systems
Organised by: ELSA and ENFIELD
AI systems are increasingly used to drive decision making and control, including in security- and safety-critical application domains.
The recent approval of the AI Act in EU introduces a strict regulatory framework for AI-powered systems that foresees clear obligations to assess and reduce risks, to be transparent and to ensure human oversight.
Being able to state with certain confidence that an AI system will behave within its designed limitations and will comply with such regulatory requirements is a sine-qua-non condition for the controlled deployment of AI. This becomes especially important for general-purpose AI models that can be used to power a wide range of applications, some unforeseen at the time of their development.
Following by-design principles for developing safe and secure AI is essential. But over and above being safe and secure by design, AI systems should also be “auditable by design”, providing enough insight and transparency about the underlying models and the training procedures employed to enable third parties to continuously evaluate their behaviour in a range of scenarios. In parallel, new tools and techniques should be developed to power the right AI auditing mechanisms, that must go over and above typical performance evaluation and benchmarking.
In this workshop we expect to discuss and debate on the issues above and identify ways for the networks of excellence to practically contribute towards developing meaningful AI auditing tools and procedures.
PW 2: Navigating Interdisciplinarity: Leveraging Insights from ICT-48 Projects to Enhance Synergies Across Academic Disciplines
Organised by: AI4Media
In this hands-on sessions session we collect good practices, do’s and don’ts and also identify what next steps could be done to take interdisciplinarity to the next level – think of suggestions for the Horizon program, education, but also the way research is rewarded, researchers are trained, and so on.
PW 3: EuroGen: Mapping the Future with Generative AI
Organised by: euROBIN and HumanE-AI-NET
tbc
14:00 – 15:30
PW 4: Next-generation infrastructure federation: a Virtual research lab for edge AI
Organised by: dAIEDGE
The workshop will focus on the concept of next-generation infrastructure federation, specifically highlighting the development of a virtual research lab tailored for edge AI. This lab will serve as a collaborative platform for researchers and developers to conduct experiments and research across various domains of edge AI.
Key aspects of the workshop will include discussions on facilitating common research endeavors, providing access to dedicated resources, tools, and services within a decentralized network. Attendees will explore how this infrastructure can enhance edge resource availability, optimize costs, and foster innovation in the field of AI at the edge. Through interactive session and discussions, participants will exchange views and gain insights into the potential of virtual research labs to advance edge AI technologies and address real-world challenges effectively.
PW 5: Assessing and Enhancing Fairness in AI Systems
Organised by: AI4Media and ELIAS
AI algorithms exhibit biases that are inherited from those present in their training data and these biases might lead to unfair decisions in real-life situations.
While many existing efforts focus on bias detection and understanding, this workshop will address proactive ways to counter the phenomenon. Particular attention will be given to three topics:
- Interdisciplinary initiatives that seek to make AI fairer, such as initiatives to reduce dataset biases by design.
- Effects of biases in high-impact AI applications (face recognition, recommenders, automatic scoring, media analysis).
- Representational biases in Large Multimodal and Language Models.
PW 6: NoE collaborative actions: Joint SRA and Ecosystem Mapping
Organised by: VISION, TAILOR, ELISE
Since the 1st community workshop, the EC and the Networks of Excellence (NoE) proclaimed that a mapping of the AI ecosystem is a promising, strategic topic. Together VISION and the NoEs undertook an extensive mapping and developed a prototype of the tool, and learned several lessons along the way.
Next, to ensure the further development, the NoEs have joined forces with Adra-e and the AI-on-demand in founding an ADRA topic group with the mandate to guide the development and governance of the map. In this session we will present our shared vision and invite the audience to join the discussion on what such a mapping could provide to be useful.
The six Networks of Excellence Centres in AI and Robotics (NoEs) are providing a Joint Strategic Research Agenda (SRA). Version 2 will be launched in June 2024. The European Union’s aspirations for AI, Data and Robotics (ADR) that are “made in Europe” demand an ambitious approach to advancing European AI research and development. The six Networks of Excellence Centres in AI and Robotics (NoEs) – AI4Media, ELISE, ELSA, euROBIN, HUMANE-AI-Net, and TAILOR – are providing a framework for delivering these ambitions, by advancing the frontiers of AI, data and robotics research and its translation to real-world impact in different domains.
Agenda:
- Introduction
- Joint SRA presentation & discussion
- Ecosystem mapping & brainstorming
- Group discussion about alignment
- Next steps
Thursday 27 June 2024: AIDA Symposium
9:00 – 10:00 AI University Education
Evangelos Kanoulas: AI Literacy Across the Campus
The rise of artificial intelligence (AI) technologies has started a new chapter in our world, making AI a part of almost every area of life today. This change highlights the urgent need for a wide-ranging AI education that goes beyond just computer science to cover all fields of study. In this presentation, we aim to look into how AI education is being woven into the fabric of university courses and programs. We’ll discuss why it’s important to update and broaden our approaches to learning to keep pace with the ongoing AI revolution. In this presentation, I will initially address several challenges encountered in AI education across campuses, including diverse student backgrounds, curriculum integration, and resource limitations. Subsequently, I will present the University of Amsterdam as a case study to showcase a variety of strategies devised to overcome these challenges. Topics will include bolstering computational resources, crafting an AI pedagogy that emphasizes hands-on experiential learning, fostering collaboration with industrial partners within the AI ecosystem, and incorporating industrial projects and theses. Furthermore, I will introduce a model for AI literacy tailored to the diverse educational needs and disciplines, and delineate the creation of specialized programs and pathways aimed at equipping students with vital AI skills and competencies. Additionally, the presentation will concentrate on leveraging AI to scale AI education, exploring how AI technologies themselves can be utilized to mitigate the expertise gap in AI and augment the provision of AI education across campuses.
Vincenzo Lomonaco: Artificial Intelligence Soft and Hard Skills: Current Italian Education and Exploitation Practices
Since the launch of ChatGPT, the rush has been on to build new applications of AI in education. This presentation will look at how decades of research can help inform efforts to integrate the latest large language models into digital learning environments. We will talk about the promise of AI to help solve persistent problems in education and also potential pitfalls. In doing so, we will discuss the development of Khanmigo, Khan Academy’s tutor for students and assistant for teachers.
15:00 – 16:00 AI tools in education (impact of LLMs and Generative AI in education, exams, essays)
Georgios Tsatsaronis: Artificial Intelligence in Research, Health and Life Sciences
The rapid advancements of Generative Artificial Intelligence (GAI) have created the circumstances for a plethora of opportunities in the Research space, as well as in the space of Health and Life Sciences. In parallel, unprecedented challenges await for us ahead, to ensure the development of responsible and ethical AI solutions that can serve our Research and Professional communities in the best possible way. In this talk we are going to focus on the most representative text-to-text GAI architectures that can deliver applied AI solutions with impact, leveraging the immense capabilities of large language models and other types of deep neural networks. We are also going to highlight the most important risks and challenges, and provide an overview of the latest evolutions in the research and governance space to address them. We are going to conclude the talk by demonstrating some of the latest, cutting edge GAI solutions that Elsevier has developed and released, serving the research and professional communities and supporting them in advancing science.
Kristen Dicerbo: AI in Education: A Khanmigo Case Study
Since the launch of ChatGPT, the rush has been on to build new applications of AI in education. This presentation will look at how decades of research can help inform efforts to integrate the latest large language models into digital learning environments. We will talk about the promise of AI to help solve persistent problems in education and also potential pitfalls. In doing so, we will discuss the development of Khanmigo, Khan Academy’s tutor for students and assistant for teachers.
Organising Committee:
The 4th European AI Network of Excellence Community Workshop is organised by a committee of representatives from the NoEs and relevant organizations (sorted by organization name):
- Filareti Tsalakanidou – AI4Media (local organizer)
- Ioannis Pitas – AIDA (local organizer)
- Efthymia Patmanidou – AIDA
- Candela Bravo – AI4Media
- Mariana Carola – AI4Media
- Mohamed Selim – dAIEDGE
- Cecilia Zanazzo – ELIAS
- Lorraine Wolter – ELSA
- Passant El Agroudy – HumanE-AI-Net
- Myrra Carstens – VISION
- Claudio Lazo – VISION
- Eva Dolezalova – VISION
- Joachim de Greeff – VISION
- Long Pham – VISION, AI4Europe