Přednáška Towards Ethical Artificial Intelligence: Consolidating Transparency, Fairness and Accountability in Machine Learning Models Seminář KEG

Alberto Fernández-Hilario (University of Granada, Spain)

Artificial Intelligence is being applied in a multitude of scenarios that are sensitive to the human user, i.e., medical diagnosis, granting loans, human resources management, among many others. Behind most of these Artificial Intelligence tools is a pattern recognition model generated by Machine Learning. To do this, it is necessary to start from a dataset that characterizes the problem under study, and "train" this model to represent the former information through different mathematical approximations. Thus, when sensitive applications and mathematical models are placed in the same equation, mistrust arises about the correct functioning of Artificial Intelligence systems. Among other questions raised, there is mainly what is the reason behind which the model makes one decision and not another. The answer lies in the interpretability or transparency of the model itself, i.e., that its components are directly understandable by the human user. When this is not possible, a posteriori explainability mechanisms are used to facilitate knowledge of which variables or characteristics the model has considered. Throughout this seminar, we will introduce the current trends to achieve a trustworthy Artificial Intelligence. We will expose the components that allow a model to be transparent, as well as the existing techniques to explain more complex models such as those based on Deep Learning. Finally, we will expose some prospects that can be considered to keep improving the explanations and to allow a wider use of Machine Learning solutions in all fields of application.

Dr. Alberto Fernández-Hilario obtained his PhD (funded by a Spanish FPI grant) at the University of Granada in 2010. He is currently a Full Professor at the University of Granada, specifically in the Department of Computer Science and Artificial Intelligence, and belongs to the DASCI Research Institute. His research work focuses on the areas of Data Science, Big Data, Ethics and Trustworthy Artificial Intelligence (explainability and bias in Machine Learning models), as well as solving singular classification problems, always from interdisciplinary, applied and problem-oriented approach. He has authored more than 60 papers in relevant JCR journals, and has participated in several international peer-reviewed conferences. He has received more than 10,000 citations of specialized papers according to Web of Science, reaching an H-index of 38. He has edited a monographic book entitled "Learning from Imbalanced Datasets" (2018, Springer Ed.) that accumulates more than 700 citations with more than 64,000 accesses in total. Dr. Fernandez has participated in numerous projects and public contracts, with more than 500,000€ in competitive funding. His most recent European research project was an ITN H2020 in the field of Bioinformatics. He also exercises his leadership in transfer contracts according to art. 83 of the Organic Law 6/2001, on Universities and Collaborative Contracts, during the last 3 years (2021-2023). All this indicates his involvement and knowledge of the tools, algorithms and latest research trends in these areas. He has extensive knowledge of the research needs and expectations of academia and industry, and has previous experience in engaging with stakeholders in both sectors. Throughout his professional career, he has received different recognitions, including being considered Highly Cited Researcher in the field of Computer Science in 2017; he is currently within the top 2% of the most influential researchers in the world according to the Standford ranking, specifically in the 64th/149th position in the Ranking of Computer Science in Spain.