Max-Margin Markov Networks

By B. Taskar, Carlos Guestrin, Daphne Koller. 2003

1,512 citations recorded by the source corpus.

How researchers see this paper

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

In Natural Language Processing

Citations in the corpus

Going deeper with convolutions (2014, methodology): “Recent work has focused on motion descriptors that are invariant to camera motion [5, 6, 8, 10, 11, 12, 13, 14].”
Structured Convolutional Kernel Networks for Airline Crew Scheduling (2021, methodology): “We ﬁrst sanity check the approach on the OCR dataset (Taskar et al., 2004) yielding a test accuracy comparable to state of the art.”
Quantifying Learning Guarantees for Convex but Inconsistent Surrogates (2018, background): “At the same time it is known that inconsistent surrogates often work well in practice (for example, the Crammer-Singer formulation of multi-class SVM [8], or its generalization structured SVM [20, 22]).”
Adaptive Stochastic Dual Coordinate Ascent for Conditional Random Fields (2017, methodology): “We implemented SDCA to train a classifier on four CRF training tasks: (1) the optical character recognition (OCR) dataset (Taskar et al., 2004), (2) the CoNLL-2000 shallow parse chunking dataset (CONLL), (3) the CoNLL-2002 Dutch named-entity recognition dataset (NER), and (4) a part-of-speech (POS) ”
On Structured Prediction Theory with Calibrated Convex Surrogate Losses (2017, methodology): “It is known that Crammer-Singer multi-class SVM [14], which SSVM is built on, is not consistent for 0-1 loss unless there is a majority class with probability at least 12 [45, 28].”
Frank-Wolfe Algorithms for Saddle Point Problems (2016, methodology): “We performed experiments with 100 examples from the OCR dataset (dω = 4028) (Taskar et al., 2003).”
Minding the Gaps for Block Frank-Wolfe Optimization of Structured SVMs (2016, methodology): “Figure 2 illustrates the effect of the parameters on OCR dataset (Taskar et al., 2003) and motivates their choice.”
Block-Coordinate Frank-Wolfe Optimization for Structural SVMs (2012, methodology): “Thus, the Frank-Wolfe algorithm has the same wide applicability as subgradient methods, and can be applied to problems such as low-treewidth graphical models (Taskar et al., 2003), graph matchings (Caetano et al., 2009), and associative Markov networks (Taskar, 2004).”
A Reduction of Imitation Learning and Structured Prediction to No-Regret Online Learning (2010, methodology): “(2009) in treating structured prediction as a degenerate imitation learning problem, we apply DAGGER to the OCR (Taskar et al., 2003) benchmark prediction problem achieving results competitive with the state-of-the-art (Taskar et al.”
Learning to Map Sentences to Logical Form: Structured Classification with Probabilistic Categorial Grammars (2005, background): “…this area has focused on problems such as sequence learning, where y is a sequence of state labels (e.g., see (Lafferty, McCallum, & Pereira, 2001; Taskar, Guestrin, & Koller, 2003)), or nat-ural language parsing, where y is a context-free parse tree for a sentence x (e.g., see Taskar et al.…”
Search-based structured prediction (2009, positive): “The ﬁrst, “small,” is the problem considered by [52].”
Learning Deep Structured Models (2014, methodology): “ghest scoring conﬁguration y : y = argmax y F(x;y;w): Log-linear models are a special case of this function, with F(x;y;w) = w>˚(x;y). In this setting CRFs [14], structured Support Vector Machines [27, 29] or interpolations of those [7] are typically employed to learn the linear weights w. In thi”
Fastfood: Approximate Kernel Expansions in Loglinear Time (2014, methodology): “…have proven to be a highly successful technique for solving many problems in machine learning, ranging from classiﬁcation and regression to sequence annotation and feature extraction (Boser et al., 1992; Cortes and Vapnik, 1995; Vapnik et al., 1997; Taskar et al., 2004; Sch¨olkopf et al., 1998).”
Recent Trends in Deep Learning Based Natural Language Processing (2017, methodology): “The model is trained with the max-margin objective [109].”
Deep Sets (2017, background): “…large margin ranking performance such that the score s ( Instead, we take recourse to large-margin classiﬁcation with structured loss functions (Taskar et al., 2004) to obtain the relative loss objective In other words, we want to ensure that s ( x | X ) ≥ s ( x (cid:48) | X ) + ∆( x, x…”
Understanding Image Representations by Measuring Their Equivariance and Equivalence (2014, background): “5 to the task of structuredoutput regression (Taskar et al. 2003).”
Reward Augmented Maximum Likelihood for Neural Structured Prediction (2016, methodology): “The foundations of learning structured output models were established by the seminal work on conditional random fields (CRFs) [17] and structured large margin methods [32], which demonstrate how generalization performance can be significantly improved when one considers the joint effects of the pred”