Explainable Trustworthy brain-like AI for Data Intensive Applications

Project Information

EXTRA-BRAIN

Grant agreement ID: 101135809

DOI

10.3030/101135809

EC signature date 22 November 2023

Start date 1 January 2024

End date 28 February 2027

Funded under

• Digital, Industry and Space

Total cost

€ 4 994 312,50

EU contribution

€ 4 994 312,50

Coordinated by

KUNGLIGA TEKNISKA HOEGSKOLAN
Sweden

CORDIS provides links to public deliverables and publications of HORIZON projects.

Links to deliverables and publications from FP7 projects, as well as links to some specific result types such as dataset and software, are dynamically retrieved from OpenAIRE .

Deliverables

Documents, reports (12)

Impact realisation roadmap

Document outlining the roadmap for impact

Mid-project revision of the Data Management Plan

Mid-project update (revision) of the Data Management Plan.

Requirements, use-case scenarios, specification and architecture

Report describing the collected requirements and specification for use cases

Project reference manual (quality, risks and contingency plans)

Reference manual for the project created at its start

Collaboration, clustering and standardization activities - halfway report

Document outlining collaborative and clustering efforts and standardization activities - halfway report.

Business and sustainability analysis and planning - halfway report

Document reporting halfway analysis of business and sustainability

Virtual training framework and environment for brain-like AI models - stage I

Report on the development and validation of the proposed virtual framework for training brain-like neural networks (AI models) - first stage

Ethics Advisor's report 1

First report prepared by the external independent Ethics Advisor

Design, training, validation and deployment of brain-like neural networks in flexible learning scenarios

Report on the initial development of brain-like neural networks and their validation in various flexible learning scenarios.

Data processing pipeline for data optimisation, explainability and federated learning - stage I

Report on data processing pipelines with emphasis on the data optimisation and visualisation/explainability (stage I)

Explainable, trustworthy, transparent and ethical brain-like AI methods - stage I

Report on the development of techniques for incorporating the explainability, trustworthiness and transparency criteria for data and models - introductory stage (outline)

Use-case implementation planning and evaluations - first stage

Report on implementation, execution and evaluation of use cases - first stage (focus on implementation and validation)

Ethics Requirements (1)

OEI - Requirement No. 1

Ethics Advisor_It is not fully clear how the participants of the co-design, co-creation and evaluation activities will be recruited.The actual extent of personal data collection and processing is also not sufficiently described, since it is planned to integrate SSH disciplines such religious viewpoints for understanding the application areas, phycological, legal, and sociological aspects considering intersectional factors (gender, ethnicity, age, socioeconomic status, disability) (DoA Part B, page 19), while it is also stated in the self-assessment that information regarding race, gender, religion or political orientation will never be requested. Thus, it should be explained how all of the data intended to be processed in the project is relevant and limited to the purposes of the research project (in accordance with the ‘data minimisation ‘principle).The AI to be developed/used in the project may lead to discrimination of people, potential bias, social or economic disadvantage especially in the AI for Digital Finance sector (DoA Part B, page 16) where solutions are provided despite the lack of adequate quantity of training datasets. Therefore, the ethics risks related to the development/deployment of the AI systems/techniques should be evaluated more thoroughly and mitigation plans for potential negative social impacts should be considered.An independent Ethics Advisor should be appointed with a mandate to oversee the project activities, in particular task T1.4 and provide relevant requirements with regards to participation of humans, personal data protection and AI. Reports prepared by the external independent Ethics Advisor must be submitted as deliverables in M12 and 24.The applicant is invited to send the CV of the suggested Ethics Advisor and discuss their appointment with the Project Officer as soon as possible after the start of the action. Guidance for Ethics Advisors/Boards can be found under the European Commission Funding and Tenders portal.

Data Management Plan (1)

Data management plan

Data Management Plan

Publications

Conference proceedings (3)

Beyond One-Size-Fits-All: How User Objectives Shape Counterfactual Explanations

Author(s): Orfeas Menis Mastromichalakis, Jason Liartis and Giorgos Stamou
Published in: 3rd World Conference on eXplainable Artificial Intelligence
Publisher: 3rd World Conference on eXplainable Artificial Intelligence

Embedded FPGA Acceleration of Brain-Like Neural Networks: Online Learning to Scalable Inference

Author(s): Hafiz, M. I. A., Ravichandran, N., Lansner, A., Herman, P., & Podobas, A.
Published in: 18th IEEE MCSoC 2025
Publisher: 18th IEEE MCSoC 2025

Efficient Navigation for Quadruped Robots in Post-Disaster Scenarios

Author(s): Cruz, C., Guijarro Tolón, J., del Cerro, J., Barrientos, A.
Published in: 2025 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2025
Publisher: 2025 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

Peer reviewed articles (3)

Spiking representation learning for associative memories

Author(s): Naresh Ravichandran; Anders Lansner; Anders Lansner; Pawel Herman; Pawel Herman; Pawel Herman
Published in: Frontiers in Neuroscience, 2024, ISSN 1662-453X
Publisher: Frontiers in Neuroscience
DOI: 10.48550/ARXIV.2406.03054

Scientific Reports

Author(s): N. Chrysanthidis; F. Fiebig; A. Lansner; P. Herman
Published in: Scientific Reports, 2025, ISSN 2045-2322
Publisher: Nature Publishing Group
DOI: 10.1038/S41598-025-12611-5

Unsupervised representation learning with Hebbian synaptic and structural plasticity in brain-like feedforward neural networks

Author(s): Naresh Ravichandran, Anders Lansner, Pawel Herman
Published in: Neurocomputing, Issue 626, 2025, ISSN 0925-2312
Publisher: Elsevier BV
DOI: 10.1016/J.NEUCOM.2025.129440

Book chapters (1)

A Reconfigurable Stream-Based FPGA Accelerator for Bayesian Confidence Propagation Neural Networks

Author(s): Muhammad Ihsan Al Hafiz, Naresh Ravichandran, Anders Lansner, Pawel Herman, Artur Podobas
Published in: Lecture Notes in Computer Science, Applied Reconfigurable Computing. Architectures, Tools, and Applications, 2025
Publisher: Springer Nature Switzerland
DOI: 10.1007/978-3-031-87995-1_12