Training linear SVMs in linear time

By Thorsten Joachims. 2006

2,189 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Optimization & Training, Statistical Learning Theory

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

In Optimization & Training

In Statistical Learning Theory

Citations in the corpus

Hogwild: A Lock-Free Approach to Parallelizing Stochastic Gradient Descent (2011, background): “2 , (2) and we know a priori that the examples z ↵ are very sparse (see for example [14]).”
Neural Responding Machine for Short-Text Conversation (2015, methodology): “We use ranking SVM model (Joachims, 2006) for the parameters ω i based on the labeled dataset.”
Text Categorization as a Graph Classification Problem (2015, background): “Additionally, since a document is only represented by a subset of its original terms, the number of non-zero feature values per document also decreases, which matters for SVM, even for the linear kernel, when considering the dual formulation or in the primal with more recent optimization techniques ”
Learning to Branch in Mixed Integer Programming (2016, methodology): “SVM rank is an efﬁcient open-source package that implements that algorithm with the appropriate loss function (cid:6) (Joachims 2006).”
Making Gradient Descent Optimal for Strongly Convex Stochastic Optimization (2011, methodology): “We used the same 3 binary classiﬁcation datasets ( ccat , cov1 and astro-ph ) used by (Shalev-Shwartz et al., 2011) and (Joachims, 2006), to test the performance of optimization algorithms for Support Vector Machines using linear kernels.”
Stochastic dual coordinate ascent methods for regularized loss (2012, positive): “Several authors (e.g. [10, 3]) proved a linear convergence rate for solving SVM with DCA (not necessarily stochastic).”
Learning Answer-Entailing Structures for Machine Comprehension (2015, methodology): “We use phi partial order (Joachims, 2006; Dubey et al., 2009) which has been used in previous structural ranking literature to incorporate ranking structure in the feature vector φ : where, c j ( y ) = 1 if h ∗ i is above h ij in the ranking y else − 1 .”
Orthogonal Random Features (2016, background): “With accurate kernel approximation, efficient linear classifiers can be trained in the transformed space while retaining the expressive power of nonlinear methods [10, 21].”
Latent Embeddings for Zero-Shot Classification (2016, methodology): “Among these methods, CCA [13] maximizes the correlation between these two embedding spaces, [26] learns a linear compatibility between an fMRI-based image space and the semantic space, [35] learns a deep non-linear mapping between images and tags, ConSe [25] uses the probabilities of a softmax-outpu”
Recommender Systems (2016, neutral): “For such cases, specialized methods with linear performance [283] have been designed.”
Data Classification: Algorithms and Applications (2014, methodology): “Other methods such as SVMPerf [54] also leverage the sparsity of the text domain.”
Improving maximum margin matrix factorization (2008, methodology): “However, we may extend the reasoning of (Joachims 2006) to more than binary scores to obtain an O(m log m) algorithm instead.”
Exponentiated Gradient Algorithms for Conditional Random Fields and Max-Margin Markov Networks (2008, background): “Recently, several new algorithms have been presented, along with a rate of convergence analysis (Joachims, 2006; Shalev-Shwartz et al., 2007; Teo et al., 2007; Tsochantaridis et al., 2004).”