CoConut: Co-classification with output space regularization

S. Khamis, C. Lampert, in:, Proceedings of the British Machine Vision Conference 2014, BMVA Press, 2014.

Download
OA 408.17 KB
Conference Paper | Published | English
Author
Department
Abstract
In this work we introduce a new approach to co-classification, i.e. the task of jointly classifying multiple, otherwise independent, data samples. The method we present, named CoConut, is based on the idea of adding a regularizer in the label space to encode certain priors on the resulting labelings. A regularizer that encourages labelings that are smooth across the test set, for instance, can be seen as a test-time variant of the cluster assumption, which has been proven useful at training time in semi-supervised learning. A regularizer that introduces a preference for certain class proportions can be regarded as a prior distribution on the class labels. CoConut can build on existing classifiers without making any assumptions on how they were obtained and without the need to re-train them. The use of a regularizer adds a new level of flexibility. It allows the integration of potentially new information at test time, even in other modalities than what the classifiers were trained on. We evaluate our framework on six datasets, reporting a clear performance gain in classification accuracy compared to the standard classification setup that predicts labels for each test sample separately.
Publishing Year
Date Published
2014-09-01
Proceedings Title
Proceedings of the British Machine Vision Conference 2014
Conference
BMVC: British Machine Vision Conference
Conference Location
Nottingham, UK
Conference Date
2014-09-01 – 2014-09-05
IST-REx-ID

Cite this

Khamis S, Lampert C. CoConut: Co-classification with output space regularization. In: Proceedings of the British Machine Vision Conference 2014. BMVA Press; 2014.
Khamis, S., & Lampert, C. (2014). CoConut: Co-classification with output space regularization. In Proceedings of the British Machine Vision Conference 2014. Nottingham, UK: BMVA Press.
Khamis, Sameh, and Christoph Lampert. “CoConut: Co-Classification with Output Space Regularization.” In Proceedings of the British Machine Vision Conference 2014. BMVA Press, 2014.
S. Khamis and C. Lampert, “CoConut: Co-classification with output space regularization,” in Proceedings of the British Machine Vision Conference 2014, Nottingham, UK, 2014.
Khamis S, Lampert C. 2014. CoConut: Co-classification with output space regularization. Proceedings of the British Machine Vision Conference 2014. BMVC: British Machine Vision Conference
Khamis, Sameh, and Christoph Lampert. “CoConut: Co-Classification with Output Space Regularization.” Proceedings of the British Machine Vision Conference 2014, BMVA Press, 2014.
Main File(s)
Access Level
OA Open Access
Last Uploaded
2018-12-12T10:08:23Z


Export

Marked Publications

Open Data IST Research Explorer

Search this title in

Google Scholar