How can we fool LIME and SHAP? Adversarial Attacks on Post hoc Explanation Methods

By Dylan Slack, Sophie Hilgard, Emily Jia, Sameer Singh, Himabindu Lakkaraju. 2019

70 citations recorded by the source corpus.

Subfields: Safety, Robustness & Interpretability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Safety, Robustness & Interpretability

Citations in the corpus

Toward Trustworthy AI Development: Mechanisms for Supporting Verifiable Claims (2020, background): “, the reliability of existing methods is an active area of research [85] [86] [87] [88] [89] [90] [91] [92] [93].”
D’ya Like DAGs? A Survey on Structure Learning and Causal Discovery (2021, background): “The prevalence of reports of systemic bias arising from automated decision processes is increasing, and an awareness for sources of bias is critical in undertaking fair and equitable machine learning [97, 169, 224, 256].”
Deterministic Local Interpretable Model-Agnostic Explanations for Stable Explainability (2021, unknown): “The data generated with random sampling can be signiﬁcantly different from its original source [6,35].”
Bayes-Probe: Distribution-Guided Sampling for Prediction Level Sets (2020, background): “…to understand and improve models (Zeiler & Fergus, 2014), but care must be taken not to provide misleading insights and explanations (Kindermans et al., 2019; Adebayo et al., 2018; Rudin, 2019) and to consider the risk of adversarial attacks (Slack et al., 2019; Ghorbani et al., 2019a).”
RECAST: Interactive Auditing of Automatic Toxicity Detection Models (2020, unknown): “tools, like Errudite, allow users to view hypotheses with respect to the true error distribution on the entire dataset [20]. Some techniques attempt to adversarially perturb language input to a model [21,16], with some methods using a human-in-theloop design [9]. Currently deployed models also conta”