Timezone: »
In this paper we establish a duality between boosting and SVM, and use this to derive a novel discriminant dimensionality reduction algorithm. In particular, using the multiclass formulation of boosting and SVM we note that both use a combination of mapping and linear classification to maximize the multiclass margin. In SVM this is implemented using a pre-defined mapping (induced by the kernel) and optimizing the linear classifiers. In boosting the linear classifiers are pre-defined and the mapping (predictor) is learned through combination of weak learners. We argue that the intermediate mapping, e.g. boosting predictor, is preserving the discriminant aspects of the data and by controlling the dimension of this mapping it is possible to achieve discriminant low dimensional representations for the data. We use the aforementioned duality and propose a new method, Large Margin Discriminant Dimensionality Reduction (LADDER) that jointly learns the mapping and the linear classifiers in an efficient manner. This leads to a data-driven mapping which can embed data into any number of dimensions. Experimental results show that this embedding can significantly improve performance on tasks such as hashing and image/scene classification.
Author Information
Ehsan Saberian (Netflix)
Jose Costa Pereira (UC San Diego)
Nuno Nvasconcelos (UC San Diego)
Can Xu (Google)
More from the Same Authors
-
2020 Poster: Learning Representations from Audio-Visual Spatial Alignment »
Pedro Morgado · Yi Li · Nuno Nvasconcelos -
2020 Poster: Contrastive Learning with Adversarial Examples »
Chih-Hui Ho · Nuno Nvasconcelos -
2019 Poster: Deliberative Explanations: visualizing network insecurities »
Pei Wang · Nuno Nvasconcelos -
2018 Poster: Self-Supervised Generation of Spatial Audio for 360° Video »
Pedro Morgado · Nuno Nvasconcelos · Timothy Langlois · Oliver Wang -
2014 Poster: Multi-Resolution Cascades for Multiclass Object Detection »
Ehsan Saberian · Nuno Vasconcelos -
2011 Poster: Maximum Covariance Unfolding : Manifold Learning for Bimodal Data »
Vijay Mahadevan · Chi Wah Wong · Jose Costa Pereira · Tom Liu · Nuno Vasconcelos · Lawrence Saul