active learning with confidence

... slow and difficult. One popular solution is Active Learning, which maximizes learning accuracy while minimiz- ing labeling efforts. In active learning, the learning algorithm itself selects unlabeled ... margin it has low confidence compared to the second. Both the margin value and confidence should be considered in choosing which example to label. We present active learning with confidence us- ing ... important property for inter- active learning. Experimental validation on a num- ber of datasets shows that active learning with con- fidence can improve standard methods. 2 Confidence- Weighted Linear...

Ngày tải lên: 17/03/2014, 02:20

... that using confidence- weighted classifiers with transition-based dependency pars- ing yields results comparable with the state-of-the- art results achieved with Support Vector Machines - with faster ... deals with this by having a confidence- parameter for each weight, modeled by a Gaussian distribution, and this pa- rameter is used to make more aggressive updates on weights with lower confidence ... ex- periments with using CW-classifiers in transition- based parsing. 5.1 Online classifiers We compare CW-classifiers with other online al- gorithms for linear classification. We compare with perceptron...

Ngày tải lên: 20/02/2014, 04:20

Ngày tải lên: 17/03/2014, 08:20

Ngày tải lên: 22/06/2014, 06:20

... each adaptation iteration. The adaptation process using active learning is represented by the curve a, while applying count-merging with active learning is rep- resented by the curve a-c. Note that ... al. (2006), where active learning was used successfully to reduce the annotation effort for WSD of 5 English verbs using coarse-grained evaluation. In that work, the authors only used active learning ... data from WSJ. In the baseline approach, the additional WSJ examples are randomly selected. With active learning (Lewis and Gale, 1994), we use uncertainty sampling as shown D T ← the set of BC...

Ngày tải lên: 08/03/2014, 02:21

... re- ranking with the proposed discriminative language mod- els. (DLM 0) or factored features by standard percep- tron (DLM 1), confidence- weighted learning (DLM 2) and confidence- weighted learning with ... four discriminative learning algorithms, we see that fac- tored features are slightly better then POS features, confidence- weighted learning is slightly better than perceptron, and confidence- weighted learning with soft ... trained dis- criminative language models by standard percep- tron, confidence- weighted learning and confidence- weighted learning with soft margin. We then trained the weights of a re-ranker using eight...

Ngày tải lên: 17/03/2014, 00:20

Ngày tải lên: 23/03/2014, 17:20

... nhân người học, đảm bảo cho họ thích ứng với đời sống xà hội. TECHNIQUES THAT SUPPORT ACTIVE LEARNING  BRAINSTORMING  Free writing  Listing/bulleting  Clustering/mapping/webbing ... thiểu vai trò của người Khác Quá phục tùng Tự biến mình thành người vô hình Thờ ơ Active learning Một số biểu hiện cụ thể của người học theo hướng tích cực Người học biết làm chủ...

Ngày tải lên: 14/10/2013, 07:11

... directions. 2 Active Learning 2.1 Pool-based Active Learning Our base framework of active learning is based on the algorithm of (Lewis and Gale, 1994), which is called pool-based active learning. ... propose methods of improving active learning for parsing by using a smaller constituent than a sentence as a unit that is selected at each iteration of active learning. Typically in active learning for parsing ... framework of active learning, since the selection strategy with large mar- gin classifiers (Section 2.2) is much simpler and seems more practical for active learning in Japanese dependency parsing with...

Ngày tải lên: 20/02/2014, 04:20

... machine active learning with applications to text classification. Journal of Machine Learning Re- search (JMLR), 2:45–66. David Vickrey, Oscar Kipersztok, and Daphne Koller. 2010. An active learning ... Vijay-Shanker. 2009b. Tak- ing into account the differences between actively and passively acquired data: The case of active learning with support vector machines for imbal- anced datasets. In Proceedings ... Association for Computational Linguistics. Greg Schohn and David Cohn. 2000. Less is more: Active learning with support vector machines. In Proc. 17th International Conf. on Machine Learn- ing, pages...

Ngày tải lên: 20/02/2014, 04:20

... respectively. 5 Improving MT quality with N-best list reranking Experiments reporting in Section 4 indicate that the proposed confidence measure has a high correlation with HTER. However, it is not very ... hypotheses and for each hypothesis we compute sentence-level confidence scores. The best candidate is the hypothesis with high- est confidence score. Table 3 shows the performance of reranking systems ... word in MT out- put, namely Good/Bad with a binary classifier and Good/Insertion/Substitution/Shift with a 4-class classi- fier. Each classifier is trained with different feature sets as follow: •...

Ngày tải lên: 20/02/2014, 04:20

... 2002), scientific text is annotated with POS tags, parse trees, and named entities. In this paper, we introduce multi-task active learning (MTAL), an active learning paradigm for multiple annotation ... Breiman. 1996. Bagging predictors. Machine Learning, 24(2):123–140. David A. Cohn, Zoubin Ghahramani, and Michael I. Jor- dan. 1996. Active learning with statistical models. Journal of Artificial ... is extremely labor- intensive. The Active Learning (AL) paradigm (Cohn et al., 1996) offers a promising solution to deal with this bottleneck, by allowing the learning algorithm to control the...

Ngày tải lên: 20/02/2014, 09:20

... 2000, Learning to construct knowledge bases from the World Wide Web, Artificial Intelligence, 118(1-2), pp. 69-113. T. Joachims, 1998, Text Categorization with Support Vector Machines: Learning ... features. Since elements with a high TF-IDF value in projections of a feature must become more useful classification criteria for the feature, we use only elements with TF-IDF values above ... Using TCFP with those Using other Classifiers In this section, we prove the superiority of TCFP over the other classifiers (SVM, kNN, Naive Bayes (NB), Roccio) in training data with much noisy...

Ngày tải lên: 20/02/2014, 16:20

... Features that were significantly correlated with expected performance at confidence level of 0.95 are marked with (*). Overall, better perfor- mance is associated with smaller inputs, lower en- tropy, ... show the SVM classification results when inputs were paired only with those within the same year. Next inputs of all years were paired with no restrictions. We report the classifi- cation accuracies ... pages 299–306. M. Dredze and K. Czuba. 2007. Learning to admit you’re wrong: Statistical tools for evaluating web qa. In NIPS Workshop on Machine Learning for Web Search. M. Kaisser, M. A. Hearst,...

Ngày tải lên: 22/02/2014, 02:20

... Valencia 46021 Valencia, Spain fcn@dsic.upv.es Abstract This work deals with the application of confidence measures within an interactive- predictive machine translation system in order to reduce human ... Im- proving interactive machine translation via mouse actions. In Proc. EMNLP, pages 25–27. N. Ueffing and H. Ney. 2005. Application of word- level confidence measures in interactive statistical machine ... P. Plamondon. 1997. Target- text mediated interactive machine translation. Ma- chine Translation, 12:12–175. S. Gandrabur and G. Foster. 2003. Confidence esti- mation for text prediction. In Proc....

Ngày tải lên: 07/03/2014, 22:20

Ngày tải lên: 08/03/2014, 16:20

... partici- pants are presented with a ‘gold standard’ human utterance from our dataset, which they must com- pare with utterances generated by models trained with and without active learning on a set of ... through active learning, in which the next semantic input to annotate is de- termined by the current model. The probabilis- tic nature of BAGEL allows the use of certainty- based active learning ... same dialogue act can only be queried twice during the active learning procedure. A consequence is that the training set used for active learning converges towards the randomly sampled set as...

Ngày tải lên: 17/03/2014, 00:20

... word confidence estimation. (Quirk, 2004) trained a sentence level confidence mea- sure using a human annotated corpus. (Bach et al., 2008) used the sentence-pair confidence scores estimated with ... sentence pairs with word alignment, from which we obtained phrase translation pairs. We extract phrase translation tables from the baseline MaxEnt word alignment as well as the alignment with confidence- based ... 2: Correlation between sentence alignment confidence measure and F-score. measure suggests the possibility of selecting the alignment with the highest confidence score to ob- tain better alignments....

Ngày tải lên: 17/03/2014, 01:20

Ngày tải lên: 17/03/2014, 04:20