Timezone: »
Poster
On the Algorithmics and Applications of a Mixed-norm based Kernel Learning Formulation
Saketha Nath Jagarlapudi · dinesh govindaraj · Raman S · Chiranjib Bhattacharyya · Aharon Ben-Tal · K. R. Ramakrishnan
Motivated from real world problems, like object categorization, we study a particular mixed-norm regularization for Multiple Kernel Learning (MKL). It is assumed that the given set of kernels are grouped into distinct components where each component is crucial for the learning task at hand. The formulation hence employs $l_\infty$ regularization for promoting combinations at the component level and $l_1$ regularization for promoting sparsity among kernels in each component. While previous attempts have formulated this as a non-convex problem, the formulation given here is an instance of non-smooth convex optimization problem which admits an efficient Mirror-Descent (MD) based procedure. The MD procedure optimizes over product of simplexes, which is not a well-studied case in literature. Results on real-world datasets show that the new MKL formulation is well-suited for object categorization tasks and that the MD based algorithm outperforms state-of-the-art MKL solvers like \texttt{simpleMKL} in terms of computational effort.
Author Information
Saketha Nath Jagarlapudi (Indian Institute of Technology, Bombay)
dinesh govindaraj (Indian Institute of Science)
Raman S
Chiranjib Bhattacharyya (Indian Institute of Science)
Aharon Ben-Tal
K. R. Ramakrishnan (Indian Institute of Science)
More from the Same Authors
-
2015 Poster: Weighted Theta Functions and Embeddings with Applications to Max-Cut, Clustering and Summarization »
Fredrik D Johansson · Ankani Chattoraj · Chiranjib Bhattacharyya · Devdatt Dubhashi -
2015 Poster: Spectral Norm Regularization of Orthonormal Representations for Graph Transduction »
Rakesh Shivanna · Bibaswan K Chatterjee · Raman Sankaran · Chiranjib Bhattacharyya · Francis Bach -
2014 Poster: Learning on graphs using Orthonormal Representation is Statistically Consistent »
Rakesh Shivanna · Chiranjib Bhattacharyya -
2014 Poster: A provable SVD-based algorithm for learning topics in dominant admixture corpus »
Trapit Bansal · Chiranjib Bhattacharyya · Ravindran Kannan -
2012 Poster: The Lovasz $\theta$ function, SVMs and finding large dense subgraphs »
Vinay Jethava · Anders Martinsson · Chiranjib Bhattacharyya · Devdatt Dubhashi -
2010 Poster: Efficient algorithms for learning kernels from multiple similarity matrices with general convex loss functions »
Achintya Kundu · Vikram Mukunda Rao Tankasali · Chiranjib Bhattacharyya · Aharon Ben-Tal -
2007 Poster: Kernels on Attributed Pointsets with Applications »
Mehul Parsana · Sourangshu Bhattacharya · Chiranjib Bhattacharyya · K. R. Ramakrishnan -
2007 Poster: A Randomized Algorithm for Large Scale Support Vector Learning »
Krishnan S Kumar · Chiranjib Bhattacharyya · Ramesh Hariharan