ANDERSON ET AL 2008

machine translation (SMT) models is a parallelcorpus. In many cases, the same information isavailable in multiple languages simultaneously asa multilingual parallel corpus, e.g., European Par-liament (EuroParl) and U.N. proceedings. In thispaper, we consider how to use active learning (AL)in[r]

YẾU TỐ QUYẾT ĐỊNH CỦA CHÍNH SÁCH CỔ TỨC: TRƯỜNG HỢP VIỆT NAM

Có một số nhà nghiên cứu đã nghiên cứu chính sách cổ tức tại các nước phát triển như Hoa Kỳ (Chang và Rhee, 1990), Vương quốc Anh (AlNajjar và Hussainey, 2009), Argentina (Bebczuk, 2005), Ba Lan (Kowalewski et al., 2008) hay Nhật Bản (Ho,[r]

associated with each word, often a vector. Eachdimension’s value corresponds to a feature andmight even have a semantic or grammaticalinterpretation, so we call it a word feature.Conventionally, supervised lexicalized NLP ap-proaches take a word and convert it to a symbolicID, which is then transfor[r]

respectively. CorpusIT includes academic papers of 6.64M in size from Chinese IT journals be-tween 1998 and 2000. CorpusLegal includes the complete set of official Chinese constitutional law articles and Economics/Finance law articles of 1.04M in size (http://www.law-lib.com/). For comparison to pre[r]

cus solely on phrase correspondences. More-over, various methods using syntactic analyt-ical tools(Pozar and Charniak, 2006; Muttonet al., 2007; Mehay and Brew, 2007) are pro-posed to address the sentence structure. Nev-ertheless, those methods depend strongly onthe quality of the syntactic a[r]

this model only determines how good a candidateantecedent is relative to the active NP, but not howgood a candidate antecedent is relative to other can-didates. So, it fails to answer the critical question ofwhich candidate antecedent is most probable. Sec-ond, it has limitations in its expressivene[r]

ers (Galley et al., 2004). In fact, synchronous gram-mars and tree transducers can be seen as instances ofthe same more general class of automata (Shieber,3The addition of W symbols is a convenience; it is easier todesign transducer rules where every substring on the right sidecorrespo[r]

movements, walking trajectories) and environmental ones (ambient temperature, relative humidity of the air, noise level, luminosity, etc.). Recorded data are gathered and transmitted to a master computer which processes them while taking into account knowledge about the user. The software used must[r]

2006). The active form of IL-18 induces signal transduction by binding to itsreceptor, IL-18α/β receptor (IL1Rrp/IL1RAPL) expressed by diverse cell types,including neurons and glia cells. In adult brains of untreated BALB/c mice, IL-18 isconstitutively the most highly expressed cytokine (Fig. 1). In[r]

parsing accuracy usually goes down dramati-cally with the increase of sentence length, translating long sentences often takes long time and only produces degenerate transla-tions. We propose a new method named sub-sentence division that reduces the decoding time and improves the translation quality[r]

quence and dependencies between variables, in-cluding the DDAs. Bui et al. (2009) were espe-cially interested in whether DGMs better exploitnon-lexical features. Fern´andez et al. (2008) ob-tained much more value from lexical than non-lexical features (and indeed n[r]

are turning to expert and novice dancers to help address suchquestions (Calvo-Merino et al., 2005, 2006; Cross et al., 2006,2009b; Bläsing et al., 2010). One consistent finding this researchhas revealed is that when dancers observe a type of style of move-men[r]

Hoffman, 2005). The PolyU (Wenjie et al., 2008) system determines the sentiment orienta-tion with two estimated language models for the positive versus negative categories. The QUANTA (Li, 2008) system detects the opinion holder, the object and the polarity of the opinion[r]

COLLI A, ET AL.[r]

reference core relations. They noted that core relations do not contain all information provided by closed graphs. Hence their measure is only an approximation of what should be assessed. Consider the previous example again. If we are evaluating graph S2, they will fail to verify that B<D is[r]

tems are estimated to best reflect the character-istics of the training data, at the cost of porta-bility: a system will be successful only as longas the training material resembles the input thatthe model gets. Therefore, whenever we have ac-cess to a large amount of labeled data from some“source” ([r]

processing and machine learning communities. Typical works include Miller et al (2000), Ze-lenko et al (2003), Culotta and Sorensen (2004), Bunescu and Mooney (2005a), Bunescu and Mooney (2005b), Zhang et al (2005), Roth and Yih (2002), Kambhatla (2004), Zha[r]

17th edition of Harrison’s Principles of Internal Medicine. It isdesigned for the student of medicine to reinforce the knowledge contained in the parent book in an active, rather thanpassive, format. This book contains over 1000 questions,most centered on a patient presentation. Answering thequestio[r]

considerationsandlong-termfollow-up.BJUInt2002; 89:339-343.4. Kavoussi et al., 1992. Kavoussi LR, Albala DM, Basler JW, et al: Percutaneousmanagement of urolithiasis during pregnancy. J Urol 1992; 148(pt 2):1069-1071.5. Harmon et al., 1997. Harmon WJ, Sersho[r]

databases are structured data sources that are oftenemployed in answering definitional questions (e.g.,“What is X?”, “Who is X?”). The top-performingdefinitional systems at TREC (Xu et al., 2003)extract kernel facts similar question profiles builtusing structured and semi-structured resou[r]