Engineering
[Closed]
Work description
Several interpretability methods were proposed for deep learning methods, consisting of saliency maps, natural language descriptions, and rule-based and case-based explanations. From these, case-based explanations arise as one of the most intuitive for human beings, as learning by example is our natural way of reasoning. Nonetheless, case-based explanations are sometimes prohibited due to privacy issues. In applications where there is a person exposed in the image, particularly, when those images are acquired for sensitive purposes, as is the case of medical images, the use of case-based explanations is completely inhibited. Therefore, in order to use the intuitive case-based explanations to justify and understand the deep learning model's behavior, one should be able to wash away the identity before presenting those cases to the consumer of the explanations. In this project, we intend to promote a causal design for the generation of privacy-preserving case-based explanations, starting from the explicit disentanglement between medical and identity features and moving towards a causal model in which the interventions are produced in terms of high-level semantic features.
Minimum profile required
Knowledge about Machine Learning or Computer Vision.
Preference factors
Experience in research projects, and writing of scientific papers.
Application Period
Since 14 Sep 2023 to 27 Sep 2023
[Closed]
Centre
Telecommunications and Multimedia