Confidence-weighted linear classification

By Mark Dredze, Koby Crammer, Fernando C Pereira. 2008

440 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 Artificial Intelligence

Citations in the corpus

Multiple-Source Adaptation with Domain Classifiers (2020, background): “…are available for some source domains such as TVs, laptops and CD players, each previously trained on labeled data, but no labeled data or predictor is at hand for the broader category of electronics, which can be viewed as a mixture of the sub-domains [Blitzer et al., 2007, Dredze et al., 2008].”
Algorithms and Theory for Multiple-Source Adaptation (2018, background): “Finally, in sentiment analysis, accurate predictors may be available for sub-domains such as TVs, laptops and CD players, each previously trained on labeled data, but no labeled data or predictor may be at the learner’s disposal for the more general category of electronics, which can be modeled as a”
Adaptive Bound Optimization for Online Convex Optimization (2010, methodology): “Perhaps the closest algorithms in spirit to our diagonal adaptation algorithm are confidence-weighted linear classification (Drezde et al., 2008) and AROW (Crammer et al.”
Rectified Factor Networks (2015, background): “This Kullback-Leibler divergence is convex in the mean vector μq and the covariance matrix Σq of Q, simultaneously [35, 36].”
Predicting Failures of Vision Systems (2014, background): “Applications such as spam-filtering [9], natural language processing [2, 12], speech [27] and even computer vision [61] have leveraged these ideas.”
PhishDef: URL names say it all (2010, positive): “2) Confidence Weighted (CW): CW is a linear binary classification algorithm recently introduced by Dredze et al. [ 10 ].”
Feature Weighting for Improved Classifier Robustness (2009, methodology): “The algorithm requires only O(md) operations for one epoch of training when the covariance matrix is restricted to a diagonal matrix; the setting has been proven to be very efficient and effective in the NLP tasks presented in [7].”
Online multimodal deep similarity learning with application to image retrieval (2013, background): “In literature, a variety of algorithm [6, 37, 9, 11, 52, 46] have been proposed for online learning.”
Learning to Detect Phishing Webpages (2014, methodology): “To conduct all the experiments, we used WEKA (Waikato Environment for Knowledge Analysis) [25] and CW libraries [19, 22].”
Portfolio Selection : A Survey (2014, positive): “…to further exploit the second-order portfolio information, which refers to the variance of portfolio weights (not price or price relative), following the mean reversion trading idea via Confidence Weighted (CW) online learning [Dredze et al. 2008; Crammer et al. 2008, 2009; Dredze et al. 2010].”
A Systematic Review on Affective Computing: Emotion Models, Databases, and Recent Advances (2022, background): “The earliest textual sentiment database is Multi-domain sentiment (MDS) [64,65].”
Online Methods for Multi-Domain Learning and Adaptation (2008, methodology): “Conﬁdence-weighted (CW) linear classiﬁcation (Dredze et al., 2008), a new online algorithm, maintains a probabilistic measure of parameter conﬁ-dence, which may be useful in combining parameters from different domain distributions.”
Towards Transparent Systems: Semantic Characterization of Failure Modes (2014, background): “Applications such as spam-filtering [5], natural language processing [6, 7], speech [8] and even computer vision [9] have leveraged these ideas.”
Multi-domain spoken language understanding with transfer learning (2009, methodology): “To this end, we implemented four online learning algorithms; averaged perceptron (AP) (Collins, 2002), approximate large margin algorithm (ALMA) (Daumé and Marcu, 2005), passive aggressive (PA) (Crammer et al., 2006), and confidence weighted classification (CW) (Dredze et al., 2008).”
Identifying Personal Stories in Millions of Weblog Entries (2009, methodology): “We chose to use a confidence-weighted linear classifier ( Dredze et al., 2008 ) as our machine learning algorithm.”
Multi-domain learning by confidence-weighted parameter combination (2010, methodology): “Confidence-weighted (CW) linear classification (Dredze et al. 2008; Crammer et al. 2008) is an online learning algorithm for linear classifiers.”
Online Learning: A Comprehensive Survey (2018, methodology): “2 Confidence Weighted Learning (CW) The CW algorithm (Dredze et al., 2008) is motivated by the following observation: the frequency of occurrence of different features may differ a lot in an online learning task.”