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

... sentences, resulting in a total training set of 4,040k sentences. We believe that our result is quite promising for the future of unsupervised parsing. In putting our best f-score in table 4 into perspective, ... Conclusion: future of supervised parsing In this paper we have shown that the accuracy of unsupervised parsing under U-DOP* continues to grow when enlarging the training set with additional ... independent of whether the sentences come from the WSJ domain or the LA Times domain. Although the effect of adding LA Times data is weaker than adding WSJ data, it is noteworthy that the...

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

... 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 arm movements which will be interesting ... 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 ... of repetitions associated with them since they do not include the notion of an end point. The number of repetitions of the event can be specified as a cardinal number, as a frequency, or indirectly...

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 ... 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 codebook ... 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...

Tài liệu Báo cáo khoa học: Glucuronate, the precursor of vitamin C, is directly formed from UDP-glucuronate in liver pptx

... procedure. The ﬁnding that the effect of CoASH depends on the presence of ATP (althoughnot of other NTPs such as GTP and UTP; notshown) suggests that it is indirectly mediated via the formation of ... test whether the enzymes catalysing the forma-tion of glucuronate and glucuronate 1-phosphate werepresent in the lumen of the ER (or had their catalyticsite oriented towards the lumen of this ... glucuronidated (Fig. 6),indicating that inhibition was truly due to the presence of this substrate of glucuronidation. No such decrease in the inhibition was observed with time in the case of valproate,...

Tài liệu Báo cáo khoa học: Seeking the determinants of the elusive functions of Sco proteins pptx

... Hiswhich is far in the sequence from the CXXXC motif,i.e. in the b-hairpin present in the extended, solvent-exposed loop (Fig. 4). Therefore, the involvement in the metal-binding site of residues ... function of the CXXXCmotif of human Sco proteins could therefore be impli-cated not only in the maturation of the CuAsite of Cox2 but also in the maintenance of cellular copperhomeostasis. The involvement ... proteins in kinetoplasts and land plants aswell. In particular, determining the function of the duplicated plant proteins would be especially interest-ing, as in all plants one of the two proteins...

Tài liệu Báo cáo khoa học: On the mechanism of action of the antifungal agent propionate Propionyl-CoA inhibits glucose metabolism in Aspergillus nidulans doc

... formaintenance ( see b elow) d uring the extended g rowth times.Upon addition of acetate to the media containing glucoseand propionate, the growth rate of both strains increasedand the effect of ... of 2 lmolNADPÆmin)1Æmg protein)1.Determination of maintenance In order to calculate the amount of glucose used formaintenance, the wild-type strain A26 was used. Four100-mL aliquots of ... for the initial decrease of the oxalo-acetate concentration in t he equilibrium due to the accumulation of NADH [25]. Lineweaver–Burk diagramswere obtained b y use of the worksheet of the programEXEL98...

Tài liệu Báo cáo khoa học: "Modeling the Translation of Predicate-Argument Structure for SMT" ppt

... predicate, the semantic role of the argument, the head word and the boundarywords of the argument. On the target side, the trans-lation of the verbal predicate, the translation of the head word of the ... right.ing, we run the training toolkit in a parallel manner.4 Argument Reordering Model In this section we introduce the discriminative ar-gument reordering model, features and the trainingprocedure.4.1 ... (seedetails in Section 6.1) in our word-aligned bilingualtraining data. Then we extract all training events forverbal predicates which occur at least 10 times in the training data. A training event...

Tài liệu Báo cáo khoa học: "Analyzing the Errors of Unsupervised Learning" docx

... that the ﬁrst iteration of EM reinforces the systematic mis-takes of the supervised initializer. In the ﬁrst E-step, the posterior counts that are computed summarize the predictions of the supervised ... HMM, the input x is a sequence of wordsand the output y is the corresponding sequence of part -of- speech tags. In the PCFG, the input x is a sequence of POStags and the output y is a binary parse ... on the distance from the true θ∗for the HMM as we increase the number of examples. In the unsupervised case, we use the followingprocedure to obtain a surrogate forˆθ: initialize EMwith the...

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

... CRFs in the task of recognition of protein entities. The main difﬁculty of applying semi-CRFs toBio-NER lies in the computational cost at training1Krauthammer (2004) reported that the inter-annotatoragreement ... 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 ... Table1 shows the length distribution of entities in the training set of the shared task in 2004 JNLPBA.Formally, the computational cost of training semi-CRFs is O(KLN), where L is the upper boundlength...

Tài liệu Báo cáo khoa học: "Evaluating the Accuracy of an Unlexicalized Statistical Parser on the PARC DepBank" docx

... of the lack of fine-grained subcategorization infor-mation in the PoS tags and by the need to balancedepth of analysis against the size of the deriva-tion space. On the Susanne corpus, the geometricmean ... representing lemma-affix-tags and arcs representing named grammati-cal relations. The encoding of this mapping within the grammar is similar to that of F-structure map-ping in LFG. However, the ... actions in the LR ta-ble. The parser is passed as input the sequence of most probable lemma-affix-tags found by the tag-ger. During parsing, probabilities are assignedto subanalyses based on the the...

Xem thêm