Báo cáo khoa học: "Learning for Microblogs with Distant Supervision: Political Forecasting with Twitter" ppt

... Association for Computational Linguistics, pages 603–612, Avignon, France, April 23 - 27 2012. c 2012 Association for Computational Linguistics Learning for Microblogs with Distant Supervision: Political ... Approval performance by a lexicon. 6.3 New Baseline: Topic-Neutral Sentiment Distant supervision for sentiment analysis outper- forms that with a sentiment lexicon...

Ngày tải lên: 24/03/2014, 03:20

... dialogues for any task makes some degree of trial-and-error explo- ration necessary. We use both 4-gram and 5-gram user sim- ulations for testing and for training (i.e. train with 4-gram, test with ... annotated with speech acts and contexts by (Georgila et al., 2005b). Here the task domain is flight-booking, and the aim for the dialogue man- ager is to obtain values for the us...

Ngày tải lên: 17/03/2014, 04:20

... feature per word form token. An excerpt of the annotated data with all generated symbolic and numeric 1 features is presented in Table 1. Word forms (Wrd) – The word form tokens form the central ... k-nearest neighbour classifier, outperform decision tree or rule inducers precisely for this reason. Comparing the performance of machine learning algorithms is not straightforward, and deserve...

Ngày tải lên: 17/03/2014, 06:20

... Association for Computational Linguistics, pages 721–729, Jeju, Republic of Korea, 8-14 July 2012. c 2012 Association for Computational Linguistics Reducing Wrong Labels in Distant Supervision for Relation ... LD end if end for return LD types to distinguish the sentences that express differ- ent relations with the same dependency path, such as “ABBA was formed in Stockholm.” and...

Ngày tải lên: 23/03/2014, 14:20

... unadjusted and adjusted precision at top-n for various values of n for different BLP and MLN models. For both BLPs and MLNs, simple manual weights result in superior performance than the learned weights. ... Results for Individual Target Relations Table 3 shows the adjusted precision for each relation for instances inferred using logical de- duction, BLP-Manual-Weights and BLP-Lea...

Ngày tải lên: 30/03/2014, 17:20

... each OOV word was ill-formed or not; and (b) what the standard form was for ill-formed words, subject to the task definition outlined in Section 3.1. The total number of ill-formed words contained ... Blog corpus with t d = 1 and w d = 0.5. Clearly there is significant room for immprovements in these results. We leave the improvement of ill-formed word detection for future work, and per...

Ngày tải lên: 17/03/2014, 00:20

... is other title word with the second high similarity score with the word W. This formula means that a word with high ranking in a category has a high similarity score with the title word of ... clustering technique for text categorization. Since our method does not use any labeled data, it cannot be fairly compared with the former approaches. Therefore, we compare our metho...

Ngày tải lên: 20/02/2014, 16:20

... sub-parse. For example, see the rule 962 FORM → state , λx 1 .state(x 1 ) FORM → by area , λx 1 .λx 2 .area(x 1 ,x 2 ) FORM → smallest FORM 1 FORM 2 , λx 1 .smallest(x 2 ,(FORM 1 (x 1 ),FORM 2 (x 1 , ... we performed two abla- tion studies. First, we compared the performance of λ-WASP with and without conjunct regrouping (Section 4). Second, we compared the performance of λ-WASP wit...

Ngày tải lên: 17/03/2014, 04:20

... DNA insert with acmA gene [3] pDEL1 Ap r , pBluescript SK+ with 785-bp SacI ⁄ EcoRI fragment of acmA obtained by PCR with primers ALA-4 and REPDEL-1 This study pDEL2 Ap r , pBluescript SK+ with 554-bp ... (diluted 1 : 400) with 2% bovine serum albumin for 1 h. After three washing steps with 1 mL of NaCl ⁄ P i containing 0.1% Tween-20 (Sigma) (PBST), the cells were incubated for...

Ngày tải lên: 19/02/2014, 18:20

... models are trained with standard phrase-based Moses software (Koehn and others, 2007), with IBM M4 alignments, 4gram SRILM, lexical ordering for PubMed and distance ordering for the NIST system. ... 2006) as opti- mization subroutine for PRO, which efficiently handle all pairwise samples without the need for sampling. 4.2 Results Figures 2 and 3 show the results for PubMed and NI...

Ngày tải lên: 19/02/2014, 19:20