Efficient large-scale structured learning
Research output: Contribution to journal › Conference article › Research › peer-review
We introduce an algorithm, SVM-IS, for structured SVM learning that is computationally scalable to very large datasets and complex structural representations. We show that structured learning is at least as fast-and often much faster-than methods based on binary classification for problems such as deformable part models, object detection, and multiclass classification, while achieving accuracies that are at least as good. Our method allows problem-specific structural knowledge to be exploited for faster optimization by integrating with a user-defined importance sampling function. We demonstrate fast train times on two challenging large scale datasets for two very different problems: Image Net for multiclass classification and CUB-200-2011 for deformable part model training. Our method is shown to be 10-50 times faster than {SVM}{struct}for cost-sensitive multiclass classification while being about as fast as the fastest 1-vs-all methods for multiclass classification. For deformable part model training, it is shown to be 50-1000 times faster than methods based on SVM struct, mining hard negatives, and Pegasos-style stochastic gradient descent. Source code of our method is publicly available.
Original language | English |
---|---|
Article number | 6619080 |
Journal | Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition |
Pages (from-to) | 1806-1813 |
Number of pages | 8 |
ISSN | 1063-6919 |
DOIs | |
Publication status | Published - 2013 |
Externally published | Yes |
Event | 26th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2013 - Portland, OR, United States Duration: 23 Jun 2013 → 28 Jun 2013 |
Conference
Conference | 26th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2013 |
---|---|
Country | United States |
City | Portland, OR |
Period | 23/06/2013 → 28/06/2013 |
Sponsor | IEEE Computer Society |
- cost-sensitive SVM, deformable part models, object detection, optimization, structured learning, sub-gradient
Research areas
ID: 293218609