Helena Montenegro, researcher at INESC TEC, received the “Best Master’s Thesis” award granted by the Portuguese Association for Pattern Recognition (APRP), with the topic “A privacy-preserving framework for case-based interpretability in machine learning”.
The work, developed at INESC TEC’s Centre for Telecommunications and Multimedia (CTM) and the Faculty of Engineering of the University of Porto – supervised by Jaime Cardoso and Wilson Silva –, focuses on the anonymisation of medical images to enable their use as explanations of deep learning models in clinical contexts. The main objective of the researcher, currently a PhD student, was to develop deep learning models capable of editing medical images, in order to remove revealing elements of a patient’s identity, thus preserving the information necessary for clinical diagnosis. By protecting the privacy of patients, these models favour the use of medical images as explanations of Artificial Intelligence models, contributing to increase the dependability and acceptance of said models in real contexts.
“The award shows the importance of research in the areas of privacy and interpretability in deep learning, which were the focus of my thesis. This acknowledgement serves as motivation to keep researching and proposing innovative solutions that benefit from advances in deep learning, towards improving the quality of medical diagnosis”, mentioned Helena Montenegro.
The researcher mentioned in this news piece is associated with INESC TEC.
News, current topics, curiosities and so much more about INESC TEC and its community!
