Y. Freund. Sifting informative examples from a random source. In Advances in Neural Information Processing, pages 85--89, 1994. Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
Active Learning using Adaptive Resampling - Iyengar, Apte, Zhang (2000) (10 citations) (Correct)
.... Various explanations havebeenput forth for the classification accuracies achieved by these techniques [26, 18] Adaptive resampling methods like boosting are also useful in selecting relevant examples even though their original goal was to improve the performance of weak learning algorithms [14]. The application of boosting to selective labeling has been suggested in [14] without algorithmic details or experimental results. A related application of boosting to select a subset of labeled instances for nearest neighbor classifiers has been explored in [15] The closest related work [1] ....
.... achieved by these techniques [26, 18] Adaptive resampling methods like boosting are also useful in selecting relevant examples even though their original goal was to improve the performance of weak learning algorithms [14] The application of boosting to selective labeling has been suggested in [14] without algorithmic details or experimental results. A related application of boosting to select a subset of labeled instances for nearest neighbor classifiers has been explored in [15] The closest related work [1] combines the Query by Committee approach with bagging and boosting techniques. In ....
Y. Freund. Sifting informative examples from a random source. In Advances in Neural Information Processing, pages 85--89, 1994.
Committee-Based Sample Selection For Probabilistic Classifiers - Argamon-Engelson, Dagan (1999) (4 citations) (Correct)
....it to the training set. As examples are selected for training, they restrict the set of consistent concepts, i. e, the set of concepts that label all the training examples correctly (the version space) A simple version of QBC, which was analyzed by Freund et al. 1997) see also the summary in Freund, 1994), uses the following selection algorithm: 1. Draw an unlabeled input example at random from the probability distribution of the example space. 2. Select at random two hypotheses according to the prior probability distribution of the concept class, restricted to the set of consistent concepts. ....
....of committee based sampling also in non probabilistic contexts, where explicit modeling of information gain may be impossible. In such contexts, committee members might be generated by randomly varying some of the decisions made in the learning algorithm. Acknowledgments Discussions with Yoav Freund, Yishai Mansour, and Wray Buntine greatly enhanced this work. The first author was at Bar Ilan University while this work was performed, and was supported by the Fulbright Foundation during part of the work. ....
Freund, Y. (1994). Sifting informative examples from a random source. In Working Notes of the Workshop on Relevance, AAAI Fall Symposium Series, pp. 85--89.
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