Abstract
The same method that creates adversarial examples (AEs) to fool image-classifiers can be used to generate counterfactual explanations (CEs) that explain algorithmic decisions. This observation has led researchers to consider CEs as AEs by another name. We argue that the relationship to the true label and the tolerance with respect to proximity are two properties that formally distinguish CEs and AEs. Based on these arguments, we introduce CEs, AEs, and related concepts mathematically in a common framework. Furthermore, we show connections between current methods for generating CEs and AEs, and estimate that the fields will merge more and more as the number of common use-cases grows.
Highlights
Machine Learning (ML) is transforming industry, science, and our society
We discussed the relationship between counterfactual explanations (CEs) and adversarial examples (AEs)
We argued that the definitional difference between the two object classes consists in their relation to the true data labels and their proximity to the original data-point
Summary
Machine Learning (ML) is transforming industry, science, and our society. Today, ML algorithms can fix a date at the hairdresser (Leviathan and Matias 2018), determine a protein’s 3D shape from its amino-acid sequence (Senior et al 2020), and even write news articles (Brown et al 2020). The opaqueness problem describes the limited epistemic access humans have to the inner workings of ML algorithms, especially concerning the semantic interpretation of parameters, the learning process, and the human-predictability of ML decisions (Burrell 2016). This lack of interpretability has gained a lot of attention recently, which gave rise to the field eXplainable Artificial Intelligence (XAI; DoshiVelez and Kim 2017; Rudin 2019). The problem of adversarial attacks describes the fact that complex ML algorithms are vulnerable to deceptions (Papernot et al 2016a; Goodfellow et al 2015; Szegedy et al 2014).
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
