AIXPERT

An agentic, multi-layer, GenAI-powered backbone to make an AI system explainable, accountable, and transparent

OBJECTIVES

Develop explainable multi-agent systems and multimodal data fusion.
Use Knowledge Graphs to enhance AI decision-making.
Integrate RL from Human Feedback for adaptive learning.
Enable customizable AI agents powered by LLMs.
Monitor agent performance, accuracy, and user satisfaction in real time.
Apply human-centred design and usability testing.
Create an AI Assistant for explainability guidance and feedback.

VISION

Build AI systems that are explainable, transparent, and trustworthy.
Foster accountable, human-aligned decision support tools.
Combine multi-agent systems with foundation models and real-time feedback.
Promote ethical, adaptable, and user-centric AI technologies.

OUR ROLE

Our team leads WP2, which focuses on defining the vision and requirements for AI-agentic systems. By applying a human-centred and participatory design approach, we aim to ensure AIXPERT is grounded in real-world needs and technically robust. This includes identifying user requirements, specifying use case scenarios, and designing a scalable, modular system architecture to guide the entire project lifecycle.

PARTNERS

NO.	INSTITUTION	COUNTRY
1	Athena Research Centre (ARC)	Greece
2	Universitat de Barcelona	Spain
3	Stichting Amsterdam UMC	Netherlands
4	Kyklos Ltd	United Kingdom
5	Workable Software Single Member Private Company	Greece
6	University of Groningen (Rijksuniversiteit Groningen)	Netherlands
7	Philips Consumer Lifestyle B.V.	Netherlands
8	Sorbonne Université	France
9	Furhat Robotics AB	Sweden
10	Alphatheta Music Bulgaria EOOD	Bulgaria
11	Novelcore	Greece
12	Infinitivity Design Labs	France
13	Internet of Things Applications and Multi-Layer Development	Cyprus
14	Barcelona Supercomputing Center	Spain
15	Vector Institute	Canada
16	Martel GmbH	Switzerland