We will mostly focus on four of the principles towards reliable AI. Robustness and technical security, Privacy and Data Governance, Transparency through explainability, and diversity, non-discrimination and fairness. To achieve Responsible and Ethical AI we propose five lines of research covering explainability and robustness in black box models, privacy with federated learning and blockchain and bias mitigation.
Research lines
Explanability for improved reproducibility, auditing, certification and governance of AI
Towards the development of explainable generative models
Responsible and governable AI with Federated Learning and Blockchain as privacy and traceability mechanisms
Bias mitigation and accountability gap: from trustworthy AI to ethical machines
