DROCC: Deep Robust One-Class Classification

02/28/2020
by   Sachin Goyal, et al.
13

Classical approaches for one-class problems such as one-class SVM (Scholkopf et al., 1999) and isolation forest (Liu et al., 2008) require careful feature engineering when applied to structured domains like images. To alleviate this concern, state-of-the-art methods like DeepSVDD (Ruff et al., 2018) consider the natural alternative of minimizing a classical one-class loss applied to the learned final layer representations. However, such an approach suffers from the fundamental drawback that a representation that simply collapses all the inputs minimizes the one class loss; heuristics to mitigate collapsed representations provide limited benefits. In this work, we propose Deep Robust One Class Classification (DROCC) method that is robust to such a collapse by training the network to distinguish the training points from their perturbations, generated adversarially. DROCC is motivated by the assumption that the interesting class lies on a locally linear low dimensional manifold. Empirical evaluation demonstrates DROCC's effectiveness on two different one-class problem settings and on a range of real-world datasets across different domains - images(CIFAR and ImageNet), audio and timeseries, offering up to 20 over the state-of-the-art in anomaly detection.

READ FULL TEXT

page 4

page 11

research
01/29/2021

The Deep Radial Basis Function Data Descriptor (D-RBFDD) Network: A One-Class Neural Network for Anomaly Detection

Anomaly detection is a challenging problem in machine learning, and is e...
research
05/27/2020

Language Representation Models for Fine-Grained Sentiment Classification

Sentiment classification is a quickly advancing field of study with appl...
research
03/25/2021

Classification of Local Problems on Paths from the Perspective of Descriptive Combinatorics

We classify which local problems with inputs on oriented paths have so-c...
research
02/06/2023

Spatial Functa: Scaling Functa to ImageNet Classification and Generation

Neural fields, also known as implicit neural representations, have emerg...
research
02/18/2018

Anomaly Detection using One-Class Neural Networks

We propose a one-class neural network (OC-NN) model to detect anomalies ...
research
12/12/2020

Revisiting "Qualitatively Characterizing Neural Network Optimization Problems"

We revisit and extend the experiments of Goodfellow et al. (2014), who s...
research
02/11/2022

Paraphrasing Magritte's Observation

Contrast Sensitivity of the human visual system can be explained from ce...

Please sign up or login with your details

Forgot password? Click here to reset