Difference between revisions of "AI in Europe"

Latest revision as of 07:06, 29 October 2024

This page provides a partial snapshot of the various resources, organizations, and groups that make up the AI ecosystem in Europe. It focusses mainly on the entities that are the result of the policies of the European Commission and the associated strategies to implement these policies.

The term ecosystem is much over-used but in this case it is justified in that we see the creation of a network of self-sustaining, mutually-supporting, and mutually-dependent bodies, all concerned with the implementation of Europe's strategy for research, development, and innovation (RDI) in human-centric, sustainable, secure, inclusive, and trustworthy AI.

What appears below is effectively a directory of entities and their activities. As such, it is presented hierarchically, in the form of a tree. It is important to realize that this presentation therefore omits the key links between the various entities (with one or two exceptions). These links are essential: without them, there is no system, much less an ecosystem.

In the final section, Interconnections, I try to capture these links by indicating, very briefly, how the various entities work together.

signifies a recently added item

The AI Act

AI Act Tutorial, prepared by Mireille Hildebrandt, Radboud University, for the HumanE-AI-net European Network of Human-centered Artificial Intelligence.
Note: watch this space for a new tutorial being prepared by Giancarlo Gori, Vrije Universiteit Brussel (VUB), which will address the relevance of the final version of the AI Act for human-centric AI.

European Commission Documents

Artificial Intelligence for Europe 2018
Coordinated Plan on Artificial Intelligence 2018
Ethics Guidelines for Trustworthy AI 2019
On Artificial Intelligence - A European approach to excellence and trust 2020 White Paper
A European Strategy for Data 2020
List for Trustworthy Artificial Intelligence for self-assessment 2020
AI Watch - Artificial Intelligence in Public Services 2020
Coordinated Plan on Artificial Intelligence 2021 Review; see pp. 19-20 regarding the networks of AI excellence centres and the AI lighthouse

European Commission Services

European AI Office
European AI Alliance
AI Watch

Public Private Partnerships (PPP)

European Partnership on Artificial Intelligence, Data and Robotics, in which Adra - AI, Data and Robotics Association, asbl - is the private side of the partnership, involving

• Big Data Value Association BVDA
  
• Confederation of Laboratories for Artificial Intelligence Research in Europe CLAIRE
  
• European Laboratory for Learning and Intelligent Systems ELLIS
  
• European Association for Artificial Intelligence EurAi
  
• euRobotics aisbl euRobotics Association Internationale Sans But Lucratif
  
  • AICoR the euRobotics Topic Group on AI and Cognition in Robotics
    

Adra is supported by the Adra-e project.
SRIDA Adra's Strategic Research Innovation and Deployment Agenda

Associations

CLAIRE Confederation of Laboratories for Artificial Intelligence Research in Europe
ELLIS European Laboratory for Learning and Intelligent Systems

Networks of AI Excellence Centres

Towards a vibrant European network of AI excellence

VISION Coordination and Support Action (CSA)

VISION: Value and Impact through Synergy, Interaction & coOperation of Networks of AI Excellence Centres
Building on the success and organisation of CLAIRE and AI4EU"

• About VISION:
  • The Project
  • The Objectives
  • The Partners
  • Project Organization
  • Deliverables
  • Contacts
• Community
  • Posters by the four ICT48 Networks of AI Excellence Centres: AI4Media, ELISE, HumanE-AI-net, TAILOR
• Activities
  • International AI Doctoral Academy (AIDA)
  • Theme Development Workshops
    • AI in the Public Sector 7 & 9 September 2021
    • Future Mobility - Value of Data & Trust in AI 28 October 2021
  • First ICT-48 Community Workshop 30 June 2021
  • Vision for AI 2021 22 April 2021.
• News
• Mattermost Communication Platform
  • Mattermost Login
  • User Guide for the Mattermost Installation

Networks

AI4Media

AI4Media targets the Media, Society and Democracy

• About the network: Project Overview, Action Plan, Consortium, Associate Members, Video, Methodology in an Nutshell graphic
• Results: Observatory, Use Cases, Public Reports, Scientific Papers
• Media: Events, News and Articles, Newsletters
• Open Calls
• Mobility: Junior Fellows Exchange Program, Mobility Matchmaking
• AIDA - AI Doctoral Academy

dAIEdge

• dAIEdge A network of excellence for distributed, trustworthy, efficient, and scalable AI at the edge.

ELIAS

• ELIAS European Lighthouse of AI for Sustainability

ELISE

ELISE targets data-driven machine learning, in coordination with the ELLIS Society (European Laboratory for Learning and Intelligent Systems)

• Research: Strategic Research Agenda May 2021 (45 pages), Press release
• Collaboration: Open Calls for SMEs
• Mobility: for Experienced Researchers and Ph.D. students & post-docs
• Education: ELLIS Ph.D. and Post-doc Program
• Structure: Network Organization
• Communications: News and Events

ELSA

• ELSA European Lighthouse on Secure and Safe AI

ENFIELD

• European Lighthouse to Manifest Trustworthy and Green AI

HumanE-AI-net

HumanE-AI-net: European Network of Human-centered Artificial Intelligence

• About the network: Vision, Video, Workpackages, Deliverables

• People: My blue sky project for AI in Europe is ...]
• Partners, of which there are fifty-three
• Blog: sixteen videos of people's answers to the question "What is your opinion on Human-centred Artificial Intelligence"
• Videos: HumaneAI promotional video (one video), Debate and presentations (two videos), Perspectives on HumaneAI and blue sky projects (forty-eight videos)
• News Items
• Research: Research Roadmap
• Reports: Overview Note that these reports are from a previous related project, Humane AI preparatory action
  • D1.3 Humane AI Ethical Framework
  • D2.1 HumanE AI Concept and Research Plan
  • D4.1 Organizational concepts and legal frameworks ...
  • Deliverables
• Deliverables, i.e. official reports on the project results
  
• Events:
  • HumaneAI kickoff meeting April 10-12, Berlin.
  • HumaneAI European Commission workshop, May 17, Brussels.
  • HumaneAI Roadmap Workshop, June 17-18, Paris.
  • HumaneAI Strategy Workshop October 10-11, 2019, Hague.
  • HumaneAI Innovation Workshop October 30-31, 2019, Munich.
  • Moderated online panel on the vision and plans of the Humane AI Net project December 16, 2020, online.
  • AI and Human Memory February 23, 2021, online.
  • HumaneAI Network Micro-projects March 17, 2021, online.
  • HumaneAI delivering a one day event @ European Parliament March 24, 2021, Brussels.
  • AI Idea Prize: Show, Pitch & Match April 29, 2021, online.
  • Closing the Loop: Human and AI Collaboration, Interaction, and Partnership May 19, 2021, online.
  • Regulation: What we need to talk about when we talk about AI June 3, 2021, online.
  • Human-centred Artificial Intelligence June 20–25, 2021, Dagstuhl, Germany.
  • AI and Society Round Table June 30, 2021, online.
  • The Role of AI and Digital Twins in Decarbonising Construction, June 30, 2021, online.
  • ACAI 2021: The advanced course on AI on Human Centered AI October 11-14, 2021, Berlin, Germany and online.

• Micro-projects: Overview, Evolving Graph of Partnerships
  • Neural mechanism in human brain activity during weight lifting
  • Extracting narratives from clinical patient data
  • X5LEARN: Cross Modal, Cross Cultural, Cross Lingual, Cross Domain, and Cross Site interface for access to openly licensed educational materials
  • Intelligent Virtual Characters to Induce Positive Emotions
  • DIASER: DIAlog task oriented annotations for enhanced modeling of uSER
  • Venice
  • Combining symbolic and sub-symbolic approaches: Improving neural Question-Answering-Systems through Document Analysis for enhanced accuracy and efficiency in Human-AI interaction
  • Improving air quality in large cities using mobile phone and IoT data
  • Reasoning on Contextual Hierarchies via Answer Set Programming with Algebraic Measures
  • Collection of datasets tailored for HumanE-AI multimodal perception and modelling
  • Algorithmic bias and media effects
  • Causality and Explainability in Temporal Data
  • Validating fairness property in post-processing vs in-processing systems
  • What idea of AI? Social and public perception of AI
  • Normative behavior and extremism in facebook groups
  • Agent based modeling of the Human-AI ecosystem
  • Learning Individual Users’ Strategies for Adaptive UIs
  • Coping with the variability of human feedback during interactive learning through ensemble reinforcement learning
  • Educational Recommenders with Narratives
  • Social AI gossiping
  • Creation of stories and narrative from data of Cultural Heritage
  • Using Social Norms to counteract misinformation in online communities
  • Linking language and semantic memory for building narratives
  • Neural-symbolic integration: explainability and reasoning in KENN
  • A tale of two consensus. Building consensus in collaborative and self-interested scenarios
  • Online Deep-AUTOML
  • Uncertainty Handling in Highly Automated Driving: Beyond Data
  • Ethical chatbots
  • Connected vehicle simulation for AI-based applications
  • Socially aware interactions
  • Explainable vertigo diagnosis

TAILOR

TAILOR targets trustworthy AI through the integration of learning, optimisation, and reasoning.

• About the project: Video, Project Data, Organization, Objectives, List of Deliverables: Publicly-available Deliverables Available Now
• About the network: Connectivity Fund (apply here), Industry Sectors, Partners, Network Members, International Advisory Board
• Research:
  • WP2 Strategic Research and Innovation Roadmap (SRIR)
  • WP3 Trustworthy AI: Overview, WP3 Website, Kickoff Meeting, Open Meeting, Tasks, Partners, Coordinated Actions (also here and here but without any links to fact sheets), Calendar, Links, Contacts, FAQ, D3.3 Handbook on Trustworthy AI (Version 1)
  • WP4 Unifying AI Paradigms and Representations: Overview, WP4 Website, Description, Tasks, Workshops, Publications, External Links
  • WP5 Deciding and Learning How to Act: Overview, WP5 Website, Papers, Kickoff Workshop, Roadmap Workshop, Highlights, Trustworthy AI, Focus, Description, Tasks, Calendar, Meetings and Talks, Google Group, Links
  • WP6 Learning and Reasoning in Social Contexts: Overview, WP6 website
  • WP7 Automated AI Overview: How can we use AI "at the meta-level" ... by people with limited AI expertise? 2nd AutoAI Workshop
• Activities:
  • Network Collaboration (WP9)
  • Challenges
  • Dissemination and Outreach
  • Theme Development Workshops
  • Tailor Workshop 2020
  • Tailor Summer School 2021: 23-24 September & Report
  • First Tailor Conference 2021: 21-22 September & Report
  • International AI Doctoral Academy (AIDA)
• News Items

euROBIN

euROBIN European Excellence Network on AI-powered Robotics
The main goal of the euROBIN research agenda being pursued by this network of excellence is "cognition-enabled transferable embodied AI" (also known as cognitive robotics).

The AI Lighthouse

ELSA - European Lighthouse on Secure and Safe AI See Interconnections for more on its role.

The European AI-on-Demand Platform

The AI-on-demand (AIOD) platform supported by the AI4Europe Project.
An abbreviated assessment list to support the Responsible Development and Use of AI
Assessment list to support the Responsible Development and Use of AI

Testing and Experimentation Facilities (TEFs)

Testing and Experimentation Facilities under the Digital Europe Programme.

• TEF Workshops 2020
• TEF Workshops 2021
• FAQ document on Sectorial Testing and Experimentation Facilities (TEFs) (document to be downloaded)
• Presentation at the "Leading the Digital Decade" event of 2 June 2021 Video (6' 25")
• Big Picture: Artificial intelligence in the DIGITAL Europe programme

Digital Innovation Hubs (DIHs)

Yellow Pages of Digital Innovation Hubs, including candidate European Digital Innovation Hubs

European Digital Innovation Hubs (EDIHs)

These have yet to be launched. Candidate EDIHs are listed in the Yellow Pages of Digital Innovation Hubs. See Interconnections for more on their role.

Programmes for Education and Training

EU Code Week for schools
Elements of AI Two courses: Introduction to AI and Building AI
New Master Courses in AI
AIDA - AI Doctoral Academy
...

Interconnections

Recall that the AI ecosystem comprises the following entities.

• The European Commission
• Public Private Partnerships (PPP)
• Network of AI Excellence Centres and an associated Coordination and Support Action (CSA)
• The AI Lighthouse
• The European AI-on-Demand Platform
• Testing and experimentation facilities (TEFs); yet to be launched
• Digital Innovation Hubs (DIHs)
• European Digital Innovation Hubs (EDIHs); yet to be launched
• Programmes for Education and Training

The European Commission engages in consultative policy development and implements this policy by through a strategic plan, including the funding or co-funding of the various entities identified in the plan.

Together with the European Commission, the PPPs are formed by associations or organizations that represent the strategic interests of industry and research centres & laboratories. One of their main responsibilities is to identify the priority research, development, and innovation (RDI) topics, themes, technologies, and challenges and to capture them in an effective research roadmap or research agenda. This agenda provides input for public and private funding bodies that support RDI. The goal is to have a significant impact on European products and processes though the adoption of AI.

The Networks of AI Excellence Centres aim to increase cooperation among the best research teams in Europe so that they can join forces to tackle the scientific and technological challenges identified by the PPPs, and foster closer cooperation and synergy among research teams and industry. At present, there are four networks, each targeting a different section of the AI community, and each formed by a project that is funded by the European Commission. Support for the cohesion of these networks is provided by VISION, a Coordination and Support Action. The Excellence Centres generate the research results that provide the techniques, tools, and methodologies - collectively referred to as assets - that form the foundation for further development and innovation by industry. These assets are made available through the European AI on Demand Platform, the regional Digital Innovation Hubs, and the European Digital Innovation Hubs, when they are launched. The Network of AI Excellence Centres also creates education and training material.

The AI lighthouse for Europe will expand and deepen the Network of AI Excellence Centres by adding new centres and building an even stronger alliance between them. This alliance will share a common roadmap to support basic and applied research, bring about alignment with national AI efforts, foster innovation and investment, attract and retain AI talent in Europe, and create synergies and economies of scale.

The European AI-on-Demand Platform - known as AI4EU - is a one-stop shop for all matters AI: knowledge, know-how, technology, tools, services and experts. It hosts a catalogue of AI assets that can be used by developers to exploit AI and embed it in process and products. At time of writing, there were 90 assets listed in the catalog. This is one of the forums that the AI Excellence Centres use to make their results available to the community but any organization can request their asset to be listed. Note that AI4EU has a new website https://www.ai4europe.eu/ and the old one at https://www.ai4EU.eu/ is deprecated.

The Testing and experimentation facilities (TEFs) , anticipated in the Coordinated Plan on Artificial Intelligence and yet to be launched, will provide expertise and experience in testing mature technology in a given sector, in real or close-to-real conditions. The aim is to provide developers with an infrastructure for testing AI technology before bringing it to the market. TEFs are planned for four sectors - agri-food, manufacturing, healthcare, and smart cities - and technologies (edge AI). According to the Coordinated Plan on Artificial Intelligence 2021 Review, the edge AI TEF plays a special role in the AI ecosystem by enabling "companies of any size to test and experiment innovative edge AI components based on advanced low-power computing technologies, such as neuromorphic computing" and other "low-powered AI computing hardware".

The Digital Innovation Hubs (DIHs) operate on a similar basis to the European AI-on-Demand Platform, in that they act as a one-stop shop providing technical expertise and resources for experimentation so that companies can "test before they invest". DIHs also provide innovation services, e.g. training and advice on financing training. The link with the European AI-on-Demand Platform by exploiting the assets make available by AI-on-Demand Platform, making them accessible to business on a regional basis, and by sharing their experiences with the AI-on-Demand Platform.

The European Digital Innovation Hubs (EDIHs) represents a yet-to-be-launched second wave of DIHs across Europe with some 200 DIHs planned, at a cost of approx. 1.5 billion euros, co-funded by the European Commission and member states.

The final element of the ecosystem are the education and training programs that are essential to equipping the workforce with the requisite knowledge, know-how, and skills in AI. These programs are created and implemented by several bodies, including the AI Excellence centres, the DIHs and EDIHs, and, of course, any education and training institute in the EU. There are made available through several channels, including AI Excellence centres, DIHs, EDIHs, and the AI-on-Demand Platform.