Beyond Novelty Detection: Incongruent Events, When General and Specific Classifiers Disagree

Abstract

Unexpected stimuli are a challenge to any machine learning algorithm. Here, we identify distinct types of unexpected events when general-level and specific-level classifiers give conflicting predictions. We define a formal framework for the representation and processing of incongruent events: Starting from the notion of label hierarchy, we show how partial order on labels can be deduced from such hierarchies. For each event, we compute its probability in different ways, based on adjacent levels in the label hierarchy. An incongruent event is an event where the probability computed based on some more specific level is much smaller than the probability computed based on some more general level, leading to conflicting predictions. Algorithms are derived to detect incongruent events from different types of hierarchies, different applications, and a variety of data types. We present promising results for the detection of novel visual and audio objects, and new patterns of motion in video. We also discuss the detection of Out-Of- Vocabulary words in speech recognition, and the detection of incongruent events in a multimodal audiovisual scenario.

Temporal Relations in Videos for Unsupervised Activity Analysis

“…Since the hierarchy consists of a set of more general and more specific models, we can apply the anomaly reasoning as proposed in [19]. To this end, we first need to determine if an observation is well described by a certain node in the hierarchy.…”

Domain Anomaly Detection in Machine Perception: A System Architecture and Taxonomy

“…The idea advocated in [57], [56], [58] is to compare the outputs of weak and strong classifiers. A discrepancy in their output is flagged as incongruence.…”

“…Building on the current state of the art in detecting anomalous events [39], [50], [58], the main goal of the paper is to develop a general framework for anomaly detection. We…”

“…This framework is a major extension of the conventional anomaly detection approaches reviewed in the papers of Markou and Singh [35], [36] and encompasses the recent important contributions to the anomaly detection problem presented in [58].…”

“…The methodology uses jointly i) the concept of observation likelihood, ii) decision reject option, iii) congruence [58] of multiple (e.g. noncontextual and contextual [60], [33]) interpretations, iv) sensor data quality [34], v) and model drift to detect, identify and categorise different anomalies.…”