Large Margin Classification Using the Perceptron Algorithm

By Yoav Freund, Robert E. Schapire. 1998

How researchers see this paper

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

In Optimization & Training

Citations in the corpus

Reasoning about Quantities in Natural Language (2015, methodology): “A Semi-CRF model (Sarawagi and Cohen, 2004), trained using a structured Perceptron algorithm (Collins, 2002), with Parameter Averaging (Freund and Schapire, 1998).”
Train faster, generalize better: Stability of stochastic gradient descent (2015, neutral): “In convex optimization, model averaging is sometimes observed to lead to better empirical performance of SGM and closely replated updates such as the Perceptron [10].”
Overcoming the Lack of Parallel Data in Sentence Compression (2013, background): “…compression C ⊂ E and compression Ct ⊆ E predicted at iteration t , the parameter update vector at t+ 1 is given by wt+1 = wt + ∑ e∈C\Ct f(e)− ∑ e∈Ct\C f(e) (7) w is averaged over all the wt’s so that features whose weight fluctuated a lot during training are penalized (Freund & Shapire, 1999).”
Kernel Methods for Relation Extraction (2002, methodology): “We evaluated two kernel learning algorithms: Support Vector Machine (SVM) (Cortes and Vapnik , 1995) and Voted Perceptron (Freund and Schapire, 1999).”
Design Challenges and Misconceptions in Named Entity Recognition (2009, methodology): “Our baseline NER system uses a regularized averaged perceptron (Freund and Schapire, 1999).”
Solving large scale linear prediction problems using stochastic gradient descent algorithms (2004, methodology): “The corresponding result is stronger than the corresponding mistake bound for voted perceptron method in (Freund & Schapire, 1999) since in our case the convergence is with respect to the minimization of the loss function, which has important statistical consequences such as the consistency of the…”
Marginalized Kernels Between Labeled Graphs (2003, methodology): “For the classification algorithm, we used the voted kernel perceptron (Freund & Shapire, 1999), whose the performance is known to be comparable to SVMs.”
Online Learning and Online Convex Optimization (2012, methodology): “Freund and Schapire [17] demonstrated that using the Perceptron algorithm along with the voting online-to-batch technique [23] yields an efficient replacement for Support Vector Machines [14, 38, 42].”
Semi-Supervised Learning for Natural Language (2005, methodology): “The Averaged Perceptron algorithm can be seen as an approximation to the related Voted Perceptron algorithm, which has generalization bounds in terms of the total number of mistakes made on the training data (Freund and Schapire, 1998).”
On Bayesian bounds (2006, methodology): “This includes direct applications such as that of the leave-one-out method (Helmbold & Warmuth, 1995) to voted perceptrons ( Freund & Schapire, 1999b ), as well as indirect applications such as weighted majority (Littlestone & Warmuth, 1994) motivating adaboost (Freund & Schapire, 1997).”
Online AUC Maximization (2011, methodology): “The most well-known method is the Perceptron algorithm (Rosenblatt, 1958; Freund & Schapire, 1999).”
Syntactic Processing Using the Generalized Perceptron and Beam Search (2011, methodology): “It was motivated by the voted-perceptron algorithm ( Freund and Schapire 1999 ) and has been shown to give improved accuracy over the non-averaged perceptron on a number of tasks.”
Prediction, learning, and games (2006, methodology): “Generalization of this analysis to sequences that are not linearly separable was proposed by Freund and Schapire [114], Gentile and Warmuth [125], and Gentile [124].”
Large Scale Machine Learning (2004, methodology): “Mots - Clefs : Algorithmes d ' Apprentissage, Grandes Bases de Données , Multi Layer Perceptrons , Mixture d ' Experts , Support Vector Machines , Diviser pour Régner , Gradient Stochastique , Optimisation , Marge .”