Employing EM and Pool-Based Active Learning for Text Classification

By Andrew McCallum, Kamal Nigam. 1998

How researchers see this paper

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

Citations in the corpus

Reliable Uncertainty Estimates in Deep Neural Networks using Noise Contrastive Priors (2018, background): “Ensemble methods (Seung et al., 1992; McCallumzy and Nigamy, 1998; Freund et al., 1997) generate queries that have the greatest disagreement between a set of classiﬁers.”
A Wholistic View of Continual Learning with Deep Neural Networks: Forgotten Lessons and the Bridge to Active and Open World Learning (2020, background): “Early works have shown uncertainty sampling based active learning for logistic regression (Lewis and Gale, 1994) and neural networks (Seung et al., 1992; McCallum and Nigam, 1998) based on “query by committee”, an approach to estimate uncertainty by using an ensemble of neural networks.”
Machine learning in automated text categorization (2001, neutral): “McCallum 1998; Joachims 1997; McCallum and Nigam 1998; McCallum et al. 1998; Nigam et al. 2000; Schapire and Singer 2000].”
Text Classification from Labeled and Unlabeled Documents using EM (2000, background): “It is a multinomial (or in lan- guage modeling terms, “unigram”) model, where the classifier is a mixture of multinomials (McCallum & Nigam, 1998).”
Combining Committee-Based Semi-Supervised Learning and Active Learning (2010, methodology): “Although there are some approaches that combine semi-supervised and active learning to integrate their benefits, such as [1, 15-16], there is no work done to investigate the combination of committee-based semi-supervised learning with active learning.”
Active Imitation Learning via Reduction to I.I.D. Active Learning (2012, methodology): “In particular, we use density-weighted query-by-committee (McCallum and Nigam, 1998) in our implementation.”
Active and Semisupervised Learning for the Classification of Remote Sensing Images (2014, methodology): “Few studies in this direction have been reported in [49]–[52].”
A Holistic Approach to Undesired Content Detection in the Real World (2022, unknown): “Diversity sampling chooses samples to ensure sufficient diversity within the selection, commonly achieved by clustering unlabeled data and sampling from different clusters (Nguyen and Smeulders 2004), or by selecting representative samples (McCallum and Nigam 1998; Settles and Craven 2008).”
Closing the Loop: Fast, Interactive Semi-Supervised Annotation With Queries on Features and Instances (2011, methodology): “Combining EM with pool-based active learning was previously studied in the context of instance labeling (McCallum and Nigam, 1998), and we extend the method to our interactive scenario, which supports feature labeling as well.”
A theory of transfer learning with applications to active learning (2012, background): “Active learning is a powerful form of supervised machine learning characterized by interaction between the learning algorithm and data source during the learning process [CAL94,MN98,CCS00,TK01,NS04,BEYL04,DCB07,DC08,HC08].”
Toward Optimal Active Learning through Sampling Estimation of Error Reduction (2001, methodology): “The first set of experiments used Ken Lang's Newsgroups , containing 20,000 articles evenly divided among 20 UseNet discussion groups (McCallum & Nigam, 1998b).”
Analysis of a greedy active learning strategy (2004, background): “These query points are typically chosen from an unlabeled data set, a practice called pool-based learning [10].”
A machine learning approach to sentiment analysis in multilingual Web texts (2009, methodology): “At each iteration, one example (Seung et al. 1992, Lewis and Catlett 1994, Lewis and Gale 1994, Dagan and Engelson 1995, Freund et al. 1997, Liere and Tadepalli 1997, McCallum and Nigam 1998, Iyengar et al. 2000, Roy and McCallum 2001, Tong and Koller 2002, Baram et al. 2003, Saar-Tsechansky and P”
Adaptive Submodularity: Theory and Applications in Active Learning and Stochastic Optimization (2010, methodology): “Nowak (2009) considered the case in which the outcomes are binary, i.”
Batch mode active learning and its application to medical image classification (2006, methodology): “In the past, active learning has been successfully employed in a number of applications, including text categorization (McCallum & Nigam, 1998; Roy & McCallum, 2001; Tong & Koller, 2000), computer vision (Luo et al., 2004), and information retrieval (Shen & Zhai, 2005).”