Multi-task feature and kernel selection for SVMs

By Tony Jebara. 2004

262 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

Multi-task and Lifelong Learning of Kernels (2015, methodology): “A natural extension of representation learning approach was proposed for kernel methods in [12,13], where the authors described a method for learning a kernel that is shared between tasks as a combination of some base kernels using maximum entropy discrimination approach.”
Multi-task learning for natural language processing in the 2020s: where are we going? (2020, unknown): “…been applied to feed-forward neural networks (Caruana, 1997), decision trees (Caruana, 1997), random forests (Wang et al., 2008), Gaussian Processes (Bonilla et al., 2008), support-vector machines (Jebara, 2004) and, most recently, deep neural networks (Ruder, 2017) across a broad range of domains.”
Transfer learning for image classification with sparse prototype representations (2008, background): “Most transfer learning techniques can be broadly grouped into two categories: learning intermediate representations [19, 5, 16, 2], and learning small sets of relevant features that are shared across tasks [11, 13, 3, 1, 23].”
Feature-based transfer learning with real-world applications (2010, methodology): “Jebara [84] proposed to select features for multi-task learning with SVMs.”
Online Incremental Feature Learning with Denoising Autoencoders (2012, methodology): “Alternatively, there are several methods for jointly training the feature mapping and the classiﬁer, such as via multi-task learning [31, 6], transfer learning [2, 22, 17], non-parametric Bayesian learning [1, 10], and deep learning [16, 3, 28, 27, 20].”
Transfer learning for collaborative filtering via a rating-matrix generative model (2009, methodology): “This knowledge sharing fashion is similar to the feature-representation based multi-task/transfer learning, such as (Jebara, 2004; Argyriou et al., 2007; Raina et al., 2007).”
Exclusive Lasso for Multi-task Feature Selection (2010, methodology): “We present an eﬃcient algorithm for solving the related min-max optimization based on the subgradient descent method.”
Learning Multiple Tasks with Kernel Methods (2005, background): “See also the recent work by Jebara (2004) for related work on this direction.”
Convex multi-task feature learning (2008, methodology): “Although convex optimization methods have been derived for the simpler problem of feature selection [31], prior work on multi-task feature learning has been based on more complex optimization problems which are not convex [3, 9, 18] and, so, these methods are not guaranteed to converge to a global m”
Learning Feature Selection Dependencies in Multi-task Learning (2013, methodology): “Traditionally, methods for multi-task learning under the sparsity assumption have considered common relevant and irrelevant features among tasks [8, 10, 11, 12, 13, 14].”
Multi-task feature selection (2006, methodology): “Jebara [8] proposed a general formulation of feature selection which naturally extends to the multi-task case in the framework of Maximum Entropy Discrimination.”
$\ell_{p}-\ell_{q}$ Penalty for Sparse Linear and Sparse Multiple Kernel Multitask Learning (2011, background): “For instance, Jebara et al. [ 23 ] has introduced maximum entropy discrimination for solving such a problem.”
Inductive transfer with context-sensitive neural networks (2008, methodology): “…(Greene 2002; Zellner 1962; Breiman and Friedman 1998), Bayesian methods involving constraints such as hyper priors (Allenby and Rossi 1999; Arora et al. 1998; Bakker and Heskes 2003), and most recently kernel methods such as support vector machines (SVMs; Jebara 2004; Allenby and Rossi 2005).”
Multi-Task Learning for Analyzing and Sorting Large Databases of Sequential Data (2008, methodology): “This matrix is subsequently normalized to define a random walk on a graph, with graph-diffusion analysis [14] then performed to rank and sort the sequential data.”
A Probabilistic Model for Dirty Multi-task Feature Selection (2015, methodology): “The spike-and-slab prior is also used in (Jebara, 2004), where a multi-task feature selection method is derived using the maximum entropy discrimination formalism.”