Lewis, David and Jason Catlett. 1994. Heterogeneous uncertainty sampling for supervised learning. In Proceedings of the 11th International Conference on Machine Learning (ICML-94), pages 148--156. Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
First 50 documents Next 50
Active Learning For Spoken Language Understanding - Tur, Schapire, Hakkani-Tür (2003) (Correct)
....on which it trains, has been studied under the title of active learning. Previous work in active learning has concentrated on two approaches: certainty based methods and committeebased methods. In the certainty based methods, an initial system is trained using a small set of annotated examples [3]. Then the system examines and labels the unannotated examples and determines the certainties of its predictions on them. The k examples with the lowest certainties are then presented to the labelers for annotation. In the committee based methods, a distinct set of classifiers is also created ....
D. D. Lewis and J. Catlett, "Heterogeneous uncertainty sampling for supervised learning," in Proceedings of the ICML, 1994.
Active Learning Selection Strategies for Information Extraction - Finn, Limerick (2003) (Correct)
....ask the user to label them. We focus on so called selective sampling strategies [5] in which the learner picks an instance for the user to label from a large pool on unlabeled instances. Selective sampling techniques are generally regarded as being of two types: confidence or certainty based [10], or committee based [6] In each case, the learner has built a model using a certain number of labeled training documents, and must select the next document to be labeled with the goal of choosing the document that will give the maximum information. In confidence based approaches, the learner ....
....The first approach is to try to directly estimate the informativeness of a document using some measure of uncertainty . From information theory, the amount of information gained from labeling a document is equal to the uncertainty about that document before labeling it [10]. Most learning learning algorithm supports some method of estimating confidence on unseen documents. For example, one can invoke a set of learned rules on a document, and then compute a confidence for the document based on the training set accuracies of the rules that apply to that document. ....
D. D. Lewis and J. Catlett. Heterogeneous uncertainty sampling for supervised learning. In Proceedings of ICML-94, 11th International Conference on Machine Learning, 1994.
Index Driven Selective Sampling for CBR - Wiratunga, Craw, Massie (2003) (Correct)
.... about the statistical distribution of case feature values for nearest neighbour algorithms [8] using a committee based approach combined with expectation maximization for text classification [10] and using a probabilistic classifier that selects cases based on class uncertainty for C4.5 [7]. Increasingly, estimation and prediction techniques with roots in statistics are being applied to classifiers with resulting improved accuracy [3] Partitioning all labelled and unlabelled cases is a common approach that is employed by many active sampling techniques. Clustering cases in this ....
Lewis, D.D., Catlett, J.: Heterogeneous uncertainty sampling for supervised learning. In Machine Learning: Proceedings of the 11th International Conference, pages 148--156, New Brunswick, NJ, 1994. Morgan Kauffman
Enhancing Learning using Feature and Example Selection - Raman, Ioerger (2003) (Correct)
....methods, wrapper methods and methods that are part of the learning algorithm. Other classifications of example selection are shown in Figure 3. The filter method of example selection acts as a preprocessor to the learning algorithm. Various static sub sampling methods come in this category. Lewis and Catlett (1994) use one probabilistic classifier to select instances for training another classifier. Wrapper models for example selection are used to iteratively update the models using misclassified data like the windowing technique used for decision trees working on large training sets [Quinlan, 1983] ....
....its fields to test whether they come from the same distribution as the original database, and reports whether the current sample size is su#cient [John and Langley, 1996] The Probably Close Enough criterion (PCE) is a way of evaluating a sampling strategy. Another filter approach presented by Lewis and Catlett (1994) suggest the use of one learning algorithm to filter examples for the other. 2.3.2 Wrapper Methods The best known Wrapper method for example selection is boosting [Schapire, 1990] The boosting technique modifies the distribution of the data. This is done by assigning increasing probability to ....
Lewis, D.David and Catlett, Jason Catlett, Heterogeneous uncertainty sampling for supervised learning, In Proceedings of ICML-94, 11th International Conference on Machine Learning, pages 148--156, Morgan Kaufmann Publishers, New Brunswick, 1994.
Density-Based Multiscale Data Condensation - Mitra, Murthy, Pal (2002) (1 citation) (Correct)
....stratified sampling, and peepholing [3] have been in existence. However, naive sampling methods are not suitable for real world problems with noisy data since the performance of the algorithms may change unpredictably and significantly [3] Better performance is obtained using uncertainty sampling [4] and active learning [5] where a simple classifier queries for informative examples. The random sampling approach effectively ignores all the information present in the samples not chosen for membership in the reduced subset. An advanced condensation algorithm should include information from all ....
D.D. Lewis and J. Catlett, "Heterogeneous Uncertainty Sampling for Supervised Learning," Machine Learning: Proc. 11th Int'l Conf., pp. 148-156, 1994.
Active Learning with Multiple Views - Muslea (2002) (4 citations) (Correct)
No context found.
Lewis, David and Jason Catlett. 1994. Heterogeneous uncertainty sampling for supervised learning. In Proceedings of the 11th International Conference on Machine Learning (ICML-94), pages 148--156.
Bayesian Graphical Models for Adaptive Filtering - Zhang (2005) (Correct)
No context found.
D. Lewis and J. Catlett. Heterogeneous uncertainty sampling for supervised learning. In Proceedings of the Eleventh International Conference on Machine Learning. Morgan Kaufmann, 1998.
Active Feedback in Ad Hoc Information Retrieval - Shen, Zhai (2005) (Correct)
No context found.
D. D. Lewis and J. Catlett. Heterogeneous uncertainty sampling for supervised learning. In Proceedings of ICML 1994.
Exploration and Exploitation in Adaptive Filtering Based on.. - Zhang, Xu, Callan (2003) (2 citations) (Correct)
No context found.
Lewis, D., & Catlett, J. (1998). Heterogeneous uncertainty sampling for supervised learning. Proceedings of the Eleventh International Conference on Machine Learning. Morgan Kaufmann.
The Effect Of Small Disjuncts And - Class Distribution On (Correct)
No context found.
Lewis, D. D., & Catlett, J. (1994). Heterogeneous uncertainty sampling for supervised learning. In Proceedings of the Eleventh International Conference on Machine Learning, pp.148-156.
Optimizing Abstaining Classifiers using ROC Analysis Tadek.. - Tdek Pe'trek Pie (Correct)
No context found.
Lewis, Catlett, (1994). Heterogeneous uncertainty sampling for supervised learning. Proceedings 11th International Conference Machine Learning (pp. 48--156). Morgan Kaufmann Publishers, San Francisco, US. Bibliography
Active Learning Selection Strategies for Information Extraction - Aidan Finn Nicholas (Correct)
No context found.
D. D. Lewis and J. Catlett. Heterogeneous uncertainty sampling for supervised learning. In 11th International Conference on Machine Learning, 1994.
Optimizing Abstaining Classifiers using ROC Analysis - Pietraszek (2005) (Correct)
No context found.
Lewis, D. D., & Catlett, J. (1994). Heterogeneous uncertainty sampling for supervised learning. Proceedings of ICML-94, 11th International Conference on Machine Learning (pp. 148--156). Morgan Kaufmann Publishers, San Francisco, US.
Boosting with Prior Knowledge for Call Classification - Schapire, Rochery, Rahim.. (Correct)
No context found.
David Lewis and Jason Catlett. Heterogeneous uncertainty sampling for supervised learning. In Machine Learning: Proceedings of the Eleventh International Conference, 1994.
Stream-Based Active Learning for Data Selection in a Real - World Applicati On (Correct)
No context found.
David D. Lewis and Jason Catlett, `Heterogeneous uncertainty sampling for supervised learning', in Proceedings of ICML-94, 11th International Conference on Machine Learning, eds., William W. Cohen and Haym Hirsh, pp. 148--156, New Brunswick, US, (1994). Morgan Kaufmann Publishers, San Francisco, US.
Stream-Based Learning through Data Selection in a Road.. - Saunier, Midenet.. (Correct)
No context found.
D. D. Lewis and J. Catlett. Heterogeneous uncertainty sampling for supervised learning. In William W. Cohen and Haym Hirsh, editors, Proceedings of ICML94, 11th International Conference on Machine Learning, pages 148--156, New Brunswick, US, 1994. Morgan Kaufmann Publishers, San Francisco, US.
Stream-Based Learning through Data Selection In Road - Saunier (2004) (Correct)
No context found.
Lewis, D.D., Catlett, J.: Heterogeneous uncertainty sampling for supervised learning. In Cohen, W.W., Hirsh, H., eds.: Proceedings of ICML-94, 11th International Conference on Machine Learning, New Brunswick, US, Morgan Kaufmann Publishers, San Francisco, US (1994) 148--156
Exploiting Unlabeled Utterances For Spoken Language.. - Tur, Hakkani-Tur (2003) (Correct)
No context found.
D. D. Lewis and J. Catlett, "Heterogeneous uncertainty sampling for supervised learning," in Proceedings of the ICML, New Brunswick, NJ, July 1994.
Active Labeling For Spoken Language Understanding - Gokhan Tur Mazin (Correct)
No context found.
D. D. Lewis and J. Catlett, "Heterogeneous uncertainty sampling for supervised learning," in Proceedings of the ICML, 1994. 2792
Active And Unsupervised Learning for Automatic Speech.. - Riccardi, Hakkani-Tür (2003) (1 citation) (Correct)
No context found.
D.D. Lewis and J. Catlett, "Heterogeneous uncertainty sampling for supervised learning," in Proc. of the ####### International Conference on Machine Learning, 1994, pp. 148--156.
Active Learning for Automatic Speech Recognition - Hakkani-Tür, Riccardi, Gorin (Correct)
No context found.
D.D. Lewis and J. Catlett, "Heterogeneous uncertainty sampling for supervised learning," in Proc. of the 8C8 International Conference on Machine Learning, 1994, pp. 148--156.
Aggregating Web Services with active invocation and.. - Eddie Johnston And (2004) (Correct)
No context found.
Lewis, D.D., Catlett, J.: Heterogeneous uncertainty sampling for supervised learning. In Cohen, W.W., Hirsh, H., eds.: Proceedings of ICML-94, 11th International Conference on Machine Learning, New Brunswick, US, Morgan Kaufmann Publishers, San Francisco, US (1994) 148--156
Probabilistic Memory-based Collaborative Filtering - Yu, Schwaighofer, Tresp.. (2004) (Correct)
No context found.
D. Lewis and J. Catlett, "Heterogeneous uncertainty sampling for supervised learning," in Proceedings of the Eleventh International Conference on Machine Learning. 1994, pp. 148--156, Morgan Kaufmann.
Active Learning Selection Strategies for Information Extraction - Aidan Finn Nicholas (2003) (Correct)
No context found.
D. D. Lewis and J. Catlett. Heterogeneous uncertainty sampling for supervised learning. In 11th International Conference on Machine Learning, 1994.
Active Learning of Pixel Classification - Dmitry Rusakov Supervised (Correct)
No context found.
David D. Lewis, Jason Catlett, 1994, "Heterogeneous Uncertainty Sampling for Supervised Learning", Machine Learning Proceedings of the 11th International Conference.
First 50 documents Next 50
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC