DeepAI AI Chat
Log In Sign Up

Detecting Novelties with Empty Classes

by   Svenja Uhlemeyer, et al.
Universität München
Berlin Institute of Technology (Technische Universität Berlin)
Universität Paderborn
Bergische Universität Wuppertal

For open world applications, deep neural networks (DNNs) need to be aware of previously unseen data and adaptable to evolving environments. Furthermore, it is desirable to detect and learn novel classes which are not included in the DNNs underlying set of semantic classes in an unsupervised fashion. The method proposed in this article builds upon anomaly detection to retrieve out-of-distribution (OoD) data as candidates for new classes. We thereafter extend the DNN by k empty classes and fine-tune it on the OoD data samples. To this end, we introduce two loss functions, which 1) entice the DNN to assign OoD samples to the empty classes and 2) to minimize the inner-class feature distances between them. Thus, instead of ground truth which contains labels for the different novel classes, the DNN obtains a single OoD label together with a distance matrix, which is computed in advance. We perform several experiments for image classification and semantic segmentation, which demonstrate that a DNN can extend its own semantic space by multiple classes without having access to ground truth.


page 1

page 4

page 5

page 9

page 13


Towards Unsupervised Open World Semantic Segmentation

For the semantic segmentation of images, state-of-the-art deep neural ne...

Entropy Maximization and Meta Classification for Out-Of-Distribution Detection in Semantic Segmentation

Deep neural networks (DNNs) for the semantic segmentation of images are ...

Distance-Based Learning from Errors for Confidence Calibration

Deep neural networks (DNNs) are poorly-calibrated when trained in conven...

Differentiable Projection for Constrained Deep Learning

Deep neural networks (DNNs) have achieved extraordinary performance in s...

Unsupervised Early Exit in DNNs with Multiple Exits

Deep Neural Networks (DNNs) are generally designed as sequentially casca...

Class-Specific Semantic Reconstruction for Open Set Recognition

Open set recognition enables deep neural networks (DNNs) to identify sam...

Co-Representation Learning For Classification and Novel Class Detection via Deep Networks

Deep Neural Network (DNN) has been largely demonstrated to be effective ...