Privacy-Preserving Case-Based Explanations: Enabling Visual Interpretability by Protecting Privacy

Abstract

Deep Learning achieves state-of-the-art results in many domains, yet its black-box nature limits its application to real-world contexts. An intuitive way to improve the interpretability of Deep Learning models is by explaining their decisions with similar cases. However, case-based explanations cannot be used in contexts where the data exposes personal identity, as they may compromise the privacy of individuals. In this work, we identify the main limitations and challenges in the anonymization of casebased explanations of image data through a survey on case-based interpretability and image anonymization methods. We empirically analyze the anonymization methods in regards to their capacity to remove personally identifiable information while preserving relevant semantic properties of the data. Through this analysis, we conclude that most privacy-preserving methods are not sufficiently good to be applied to casebased explanations. To promote research on this topic, we formalize the privacy protection of visual casebased explanations as a multi-objective problem to preserve privacy, intelligibility, and relevant explanatory evidence regarding a predictive task. We empirically verify the potential of interpretability saliency maps as qualitative evaluation tools for anonymization. Finally, we identify and propose new lines of research to guide future work in the generation of privacy-preserving case-based explanations.

Translating theory into practice: assessing the privacy implications of concept-based explanations for biomedical AI

“… Montenegro et al (2021a) propose a generative model to privatize case-based explanations, as well as a way to derive counterfactual explanations. However, the authors later applied the method to glaucoma detection, revealing several drawbacks for the application in medical practice ( Montenegro et al, 2021b ; 2022 ).…”

“…However, their increasing relevance raises important questions about the impact they have on the privacy of an AI system. Most previous works investigating the interdependency of privacy and explainability focused on low-level XAI methods based on input feature attribution ( Milli et al, 2019 ; Shokri et al, 2021 ; Saifullah et al, 2022 ), while Montenegro et al (2022) investigated privacy-preserving case-based explanations. We argue that more complex XAI methods have received too little attention when it comes to an assessment of their implication on privacy.…”

DeepFixCX: Explainable privacy‐preserving image compression for medical image analysis

“…The works of Montenegro et al (2021Montenegro et al ( , 2022 create privacy-preserving case-based explanations with a generative adversarial network (GAN) that anonymize an input image by generating an image of a similar example. The anonymized images are used as explanations of an original input image because they serve as a kind of similar example.…”

“…Separate work claimed private data demands a right to an explanation, particularly in cases of closed‐source data and algorithms in industry (Kim & Routledge, 2018). Privacy in machine learning more generally is gaining attention (Liu et al, 2021), and explainable dataset anonymization has few prior works (Montenegro et al, 2022). Moreover, existing works do not consider compression.…”