A Multi-directional Meta-Learning Framework for Class-Generalizable Anomaly Detection
By: Padmaksha Roy, Lamine Mili, Almuatazbellah Boker
Potential Business Impact:
Finds new problems computers haven't seen before.
In this paper, we address the problem of class-generalizable anomaly detection, where the objective is to develop a unified model by focusing our learning on the available normal data and a small amount of anomaly data in order to detect the completely unseen anomalies, also referred to as the out-of-distribution (OOD) classes. Adding to this challenge is the fact that the anomaly data is rare and costly to label. To achieve this, we propose a multidirectional meta-learning algorithm -- at the inner level, the model aims to learn the manifold of the normal data (representation); at the outer level, the model is meta-tuned with a few anomaly samples to maximize the softmax confidence margin between the normal and anomaly samples (decision surface calibration), treating normals as in-distribution (ID) and anomalies as out-of-distribution (OOD). By iteratively repeating this process over multiple episodes of predominantly normal and a small number of anomaly samples, we realize a multidirectional meta-learning framework. This two-level optimization, enhanced by multidirectional training, enables stronger generalization to unseen anomaly classes.
Similar Papers
M3OOD: Automatic Selection of Multimodal OOD Detectors
Machine Learning (CS)
Chooses best AI for new, unseen data.
Log anomaly detection via Meta Learning and Prototypical Networks for Cross domain generalization
Machine Learning (CS)
Finds computer problems even in new systems.
Robust Anomaly Detection in Industrial Environments via Meta-Learning
CV and Pattern Recognition
Finds bad parts even with wrong labels.