Generative Perturbation Analysis for Probabilistic Black-Box Anomaly Attribution
We address the task of probabilistic anomaly attribution in the black-box regression setting, where the goal is to compute the probability distribution of the attribution score of each input variable, given an observed anomaly. The training dataset is assumed to be unavailable. This task differs from the standard XAI (explainable AI) scenario, since we wish to explain the anomalous deviation from a black-box prediction rather than the black-box model itself. We begin by showing that mainstream model-agnostic explanation methods, such as the Shapley values, are not suitable for this task because of their ``deviation-agnostic property.'' We then propose a novel framework for probabilistic anomaly attribution that allows us to not only compute attribution scores as the predictive mean but also quantify the uncertainty of those scores. This is done by considering a generative process for perturbations that counter-factually bring the observed anomalous observation back to normalcy. We introduce a variational Bayes algorithm for deriving the distributions of per variable attribution scores. To the best of our knowledge, this is the first probabilistic anomaly attribution framework that is free from being deviation-agnostic.
Code (1)
Similar Papers 제목 키워드 기반
Leveraging a Probabilistic PCA Model to Understand the Multivariate Statistical Network Monitoring Framework for Network Security Anomaly Detection
Network anomaly detection is a very relevant research area nowadays, especially due to its multiple applications in the field of network security. The boost of new models based on variational autoencoders and generative …
Anomaly DetectionPROPS: Probabilistic personalization of black-box sequence models
We present PROPS, a lightweight transfer learning mechanism for sequential data. PROPS learns probabilistic perturbations around the predictions of one or more arbitrarily complex, pre-trained black box models (such as r…
ArticlesLanguage ModelingLanguage ModellingTransfer LearningQuantile-constrained Wasserstein projections for robust interpretability of numerical and machine learning models
Robustness studies of black-box models is recognized as a necessary task for numerical models based on structural equations and predictive models learned from data. These studies must assess the model's robustness to pos…
Uncertainty QuantificationWBHT: A Generative Attention Architecture for Detecting Black Hole Anomalies in Backbone Networks
We propose the Wasserstein Black Hole Transformer (WBHT) framework for detecting black hole (BH) anomalies in communication networks. These anomalies cause packet loss without failure notifications, disrupting connectivi…
Anomaly DetectionGREAT Score: Global Robustness Evaluation of Adversarial Perturbation using Generative Models
Current studies on adversarial robustness mainly focus on aggregating local robustness results from a set of data samples to evaluate and rank different models. However, the local statistics may not well represent the tr…
Adversarial Robustness