Báo cáo khoa học: "Assessing the Costs of Sampling Methods in Active Learning for Annotation" potx

... case for measuring cost in assessing AL methods. 1 IntroductionObtaining human annotations for linguistic data islabor intensive and typically the costliest part of the acquisition of an annotated ... applicable to AL for sequential labeling in general. We make the case for measuring cost in assessing AL methods by showing that the choice of a cost function signiﬁcantly affects the choice of ALalgorithm.2 ... costlythan the baseline appear above the baseline. For aspeciﬁc accuracy value on a learning curve, the cor-responding value of the cost on the random baselineis estimated by interpolation...

Báo cáo khoa học: "Assessing the Role of Discourse References in Entailment Inference" pptx

... for inference. In another 28%,references could optionally support the inference of the hypothesis. In the remaining 28%, refer-ences did not contribute towards inference. The total number of ... generic framework for modeling semantic inference. TE reduces the inference requirements of many text understand-ing applications to the problem of determiningwhether the meaning of a given textual ... crucial for entailment sincethey included the main predicate of the hypothe-sis.6This calls for an increased focus on the reso-lution of event references.Transformations. Table 3 shows the...

Tài liệu Báo cáo khoa học: "Exploring the Use of Linguistic Features in Domain and Genre Classification" potx

... the class of the majority of the items which reached it during training. The trees were grown using recursive partitioning; the splitting criterion was reduction in deviance. Using the Gini ... grave for the LPE experiments because of the ceiling effect and the small size of the complete data set, therefore, we did not rerun the corresponding experiments. Furthermore, the number of ... examples. The corresponding training sets contain the remaining 95 documents. For RIBL, test set performance is determined using leave-one-out cross validation. Feature vectors contained either...

Báo cáo khoa học: "Evaluating the Impact of Coder Errors on Active Learning" ppt

... another label before starting the active learning trial, but retain the distribution of the different labels in the pool data (active learning with random errors); (Table 1, ALrand, 30%). In the ... margin sampling is less sensitive to cer-tain types of noise than entropy sampling (Table 2).Because of space limitations we only show curves for margin sampling. For entropy sampling, the ... trend is the same, with results being slightlylower than for margin sampling. 4 Detecting Annotation NoiseUncertainty sampling using the margin metric se-lects instances for which the difference...

Tài liệu Báo cáo khoa học: "Bucking the Trend: Large-Scale Cost-Focused Active Learning for Statistical Machine Translation" docx

... Sarkar. 2009. Active learning for multilingual statistical machine trans-lation. In Proceedings of the Joint Conference of the 47th Annual Meeting of the ACL and the 4th In- ternational Joint Conference ... for Computational Linguistics.Katrin Tomanek and Udo Hahn. 2009. Semi-supervised active learning for sequence labeling. In Proceedings of the Joint Conference of the 47th An-nual Meeting of ... translation). Figure 4shows the learning curves for the same systemsand selection methods as in Figure 3 but now the x-axis measures the number of foreign words in the training data. The difference between...

Tài liệu Báo cáo khoa học: "Using Smaller Constituents Rather Than Sentences in Active Learning for Japanese Dependency Parsing" docx

... of Japanese dependency pars-ingthat the algorithm in Figure 4 does not generateevery pair of bunsetsus.34 Active Learning for ParsingMost of the methods of active learning for parsing in ... iscalled pool-based active learning. Following theirsequential sampling algorithm, we show in Fig-ure 1 the basic ﬂow of pool-based active learning. Various methods for selecting informative exam-ples ... tool for the av-eraged perceptron in C++ and used them for ex-periments. We wrote the main program of active learning and some additional scripts in Perl and sh.6.5 Settings of Active Learning For...

Báo cáo khoa học: "Assessing the Effect of Inconsistent Assessors on Summarization Evaluation" doc

... between the original summarizerranking and the ranking after excluding topics by one ortwo worst assessors in each category.we should examine the potential impact of incon-sistent assessors on the ... too.2Therefore, it would be better to look at whether as-sessors tend to ﬁnd the same SCUs (information“nuggets”) in different summaries on the same topic,and whether they annotate them consistently. ... autoan-notated it using all remaining manually-annotatedsummaries from the same topic, and then we com-pared the resulting “autoPyramid” score with the score from the original manual annotation for thatsummary....

Tài liệu Báo cáo khoa học: "Improving the Scalability of Semi-Markov Conditional Random Fields for Named Entity Recognition" pdf

... because, in manycases, the previous label of a named entity is “O”,which indicates a non-named entity. For 98.0% of the named entities in the training data of the sharedtask in the 2004 JNLPBA, the ... than the system without it (the p-value isless than 1.0 < 10−4). The result of the precedingentity information improves the performance. On the other hand, the system with preceding infor-mation ... the final performance. In this experiment, we couldnot examine the performance without filtering us-ing all the training data, because training on all the training data without filtering required muchlarger...

Tài liệu Báo cáo khoa học: "Is the End of Supervised Parsing in Sight?" pdf

... derivations, the probability of a tree is the sum of the probabilities of the derivations producing that tree. The probability of a derivation is the product of the subtree probabilities. The original ... 402 the number of subtrees of all trees increases with the Catalan number, and only ad hoc sampling could make the method work. Since U-DOP* computes the shortest derivations (in the training ... and UML-DOP against the f-scores for U-DOP reported in Bod (2006), the CCM model in Klein and Manning (2002), the DMV dependency model in Klein and Manning (2004) and their combined model DMV+CCM....

Tài liệu Báo cáo khoa học: "Defining the Semantics of Verbal Modifiers in the Domain of Cooking Tasks" pptx

... compute the length of the intervals between stirring events. The length of a single stirring event is a default which is part of the representation of the primitive actions. The number of stirring ... system (Karlin, 1988). SEAFACT operates in the domain of cooking tasks. The domain is limited to a mini-world con- sisting of a small set of verbs chosen because they involve rather complex ... empirical, linguistic study of recipes was con- ducted with the goals of delimiting the scope of the cooking domain, identifying important verbal modifiers, and defining the semantics of those...

Xem thêm