Research On Definite Null Instantiation Recognizing

SRL has attracted much attention in recent years, as witnessed by several shared tasks in Senseval/SemEval and CoNLL conferences. The state-of-the-art in semantic role labelling has now advanced so much that a number of studies have shown that automatically inferred semantic argument structures can lead to tangible performance gains in NLP applications such as information extraction, question answering or textual entailment. However, SRL has traditionally been viewed as a sentence-internal task, hence, relations between different local semantic argument structures are disregarded and this leads to a loss of important semantic information. Therefore, it was found that the missing semantics has been particularly active role in promoting discourse understanding. Therefore, Find these missing semantic roles particularly positive significance for the promotion of the text understanding. In the FrameNet, the lack of semantic role is called the Null Instantiation, also referred to as NI, those who can establish semantic links between local semantic argument is known to Definite Null Instantiation, also referred to as DNI. The simple clues to the semantic content of the passage can be able to find by extracting the KDG of a sentence. Gap filling, getting the lexical material that is syntactically absent from the heads of the constituents that syntactically control their interpretation, is a key problem for deriving KDGs. The identification of Null Instantiated is in order to discover the Null Instantiated in sentence, and indicate which is the definite. This paper starts from the NI Task of SemEval-2010Task10, proposed a simple two-stage pipeline solution to definite null instantiation recognizing:the first stage used rule-based approach to detect null instantiations in the corpus which have been semantic roles labeled, and the second stage predicts which types the null instantiations previously detected belongs to based on maximum entropy, make an important step for DNI anaphora resolution, the main research contents as follows:(1)Research Null Instantiation detection method based on rules. Null Instantiation is caused by the lack of core frame elements, but not the core framework elements can cause a NI, this is because there is a link between the core framework three relationships between each other. In-depth study of these relationships, summed up a set of rules used to detect neutral form, followed by a NI in the form of rule-based detection experiment recall rate of60.1%.(2)Research Null Instantiation classification method based on maximum entropy. By the interpretation which received the omissions of core arguments of predicates are categorized along two dimensions:Definite Null Instantiation and Indefinite Null Instantiation, accordingly, on the base of NI detection, this paper treat NI classification as a binary classification problem, select the characteristics which most closely related to the DNI such as tgt, pos, lemma, frame etc. from the words and semantic aspects, adopted classification methods based on maximum entropy, test accuracy rate of53.5%, close to the best results of the evaluation.