Backpropagated Gradient Representations for Anomaly Detection
This addresses the challenge of distinguishing normal from abnormal data in anomaly detection, offering a computationally efficient alternative to methods like adversarial networks or autoregressive models, which require significantly more parameters.
The paper tackles the problem of anomaly detection by proposing the use of backpropagated gradients as representations, which capture model updates needed to represent data, and achieves state-of-the-art performance on benchmark image recognition datasets.
Learning representations that clearly distinguish between normal and abnormal data is key to the success of anomaly detection. Most of existing anomaly detection algorithms use activation representations from forward propagation while not exploiting gradients from backpropagation to characterize data. Gradients capture model updates required to represent data. Anomalies require more drastic model updates to fully represent them compared to normal data. Hence, we propose the utilization of backpropagated gradients as representations to characterize model behavior on anomalies and, consequently, detect such anomalies. We show that the proposed method using gradient-based representations achieves state-of-the-art anomaly detection performance in benchmark image recognition datasets. Also, we highlight the computational efficiency and the simplicity of the proposed method in comparison with other state-of-the-art methods relying on adversarial networks or autoregressive models, which require at least 27 times more model parameters than the proposed method.