Báo cáo khoa học: "Mixture Model POMDPs for Efﬁcient Handling of Uncertainty in Dialogue Management" doc

... Handling of Uncertainty in Dialogue ManagementJames HendersonUniversity of GenevaDepartment of Computer ScienceJames.Henderson@cui.unige.chOliver LemonUniversity of EdinburghSchool of Informaticsolemon@inf.ed.ac.ukAbstract In ... Proceedings of ACL-08: HLT, Short Papers (Companion Volume), pages 73–76,Columbus, Ohio, USA, June 2008.c2008 Association for Computational LinguisticsMixture Model POMDPs for Efﬁcient Handling ... Decision Processes (POMDPs) provide a formal framework for mak-ing decisions under uncertainty. Recent research in spoken dialogue systems has used POMDPs for dialogue management (Williams and Young,...

Báo cáo khoa học: "Beam-Width Prediction for Efﬁcient Context-Free Parsing" pot

... 2010.Chart pruning for fast Lexicalised-Grammar parsing. In Proceedings of the 23rd International Conferenceon Computational Linguistics, pages 1472–1479, Bei-jing, China.449Proceedings of the 49th ... parsing. Comput. Linguist.,34:513–553.Adam Pauls, Dan Klein, and Chris Quirk. 2010. Top-down k-best a* parsing. In In proceedings of the An-nual Meeting on Association for Computational Lin-guistics ... minute of processing for a sin-gle sentence of 25 words, an unacceptable amount of time for real-time applications or when process-ing millions of sentences. Deterministic algorithms for dependency...

Báo cáo khoa học: "Semi-Supervised SimHash for Efﬁcient Document Similarity Search" pptx

... ∆ implies good hashing.(a) (b)Figure 4: Computational complexity of training for dif-ferent feature dimensions for 32-bit ﬁngerprint. (a) Train-ing time (sec). (b) Training space cost (MB).the ... (Manning et al., 2002) is calculated.We also calculate the averaged intra- and inter- classHamming distance for various hashing methods. In- (a) (b)Figure 2: Precision within Hamming radius 3 for ... vector of occurrence num-bers of the terms within it. The class information of docs is considered as prior knowledge that twodocs within a same class should have more similarﬁngerprints while...

Tài liệu Báo cáo khoa học: "Translation Model Adaptation for Statistical Machine Translation with Monolingual Topic Information" doc

... easily. In this paper, we propose anovel approach for translation model adapta-tion by utilizing in- domain monolingual top-ic information instead of the in- domain bilin-gual corpora, which incorporates ... takesadvantage of both the in- domain monolingual cor-pora and the out -of- domain bilingual corpus to in- corporate the topic information into our translation model, thus breaking down the corpus barrier for translation ... train two monolingual topic models withthe following corpora: one is the source part of the out -of- domain bilingual corpus Cf out, the oth-er is the in- domain monolingual corpus Cf in in...

Tài liệu Báo cáo khoa học: "Towards Relational POMDPs for Adaptive Dialogue Management" pdf

... uncertainty, making dialogue management in such set-tings a particularly challenging problem.Traditional approaches have focused onproviding theoretical accounts for eitherthe uncertainty or ... takes as input a point in the belief spaceand outputs a set of relevant moves. For efﬁciencyreasons, this function can be precomputed of ine,by segmenting the state space into distinct regionsand ... is therefore a model for the goals orpreferences of the agent.A graphical illustration of a POMDP model asa Bayesian decision network is provided in Fig. 1. In addition, a POMDP can include...

Báo cáo khoa học: "Alignment Model Adaptation for Domain-Specific Word Alignment" pptx

... two models using the out -of- domain training data and the in- domain training data, and then build a new alignment model through linear interpolation of the two trained models. In other words, ... make use of the out -of- domain training data and the in- domain training data by interpolating the trained alignment models. One method to perform model adaptation is to directly interpolate ... center of cept i. During the training process, IBM model 3 is first trained, and then the parameters in model 3 are employed to train model 4. During the testing process, the trained model...

Tài liệu Báo cáo khoa học: "Generalized Expectation Criteria for Semi-Supervised Learning of Conditional Random Fields" pdf

... computed in a tractableamount of time, since according to the Markov as-2Often these are more complicated than picking informativefeatures as proposed in this paper. One example of the kind of operator ... eschewing constraints appliedduring inference time. Without these constraints,probabilistic models can be combined easily withone another in order to arrive at a joint model, andadding in these ... for training sequence models with constraints is Changet al. (2007). They use constraints for approximateEM training of an HMM, incorporating the con-straints by looking only at the top K...

Báo cáo khoa học: "A Descriptive Framework for Translating Speaker''''s Meaning Towards a Dialogue Translation System between Japanese and English" pot

... unification-based framework. 1. Introduction In devising a machine translation system of telephone dialogues, one of the problems is how to adequately translate tile underlying meaning of the source utterance, ... REQUESTING COMPLAINING ADVISING CONFIRMING etc. Conversely, the same intention can be conveyed through various surface expressions, as in the following variations of (2-1): REQUESTING Ill ... sentence-final particle of questions, indicate request. Kindly in (2-6) signals a request in English. In other words, even without knowledge of the context of an utterance, knowledge of communicstive...

Báo cáo khoa học: "A Formula Finder for the Automatic Synthesis of Translation Algorithms" docx

... following instructions for transcribing and encod- ing texts: Instructions for Transcribing Post-edited Texts onto Magnetic Tape (1) Explanation of Format. Machine printing appears in five ... factorization of Ts into a product of machinable algorithms applicable to situations com- monly occurring within sentences. For purposes of automatic formula finding, a specific type of factoriza- ... produced in both a readable format for human inspection and a machinable format for input to the trial translator. In- formation feedback relationships will exist between the formula finder...

Báo cáo khoa học: "Data point selection for cross-language adaptation of dependency parsers" pot

... the remaining three tree-banks as training material for each language. Thetest section of the language in question is used for testing, while the POS sequences in the target train-ing section ... selectionThe key idea in our experiments is that we can use asimple form of instance weighting, similar to what isoften used for correcting sample selection bias or for domain adaptation, to improve ... use. For example, for Bulgar-ian our baseline result using 100% of the source datais 44.5%, and the result obtained using 90% of thesource data is 70.2%. Using held-out data, we onlyuse 80% of...

Xem thêm