Zhou qiaoli: A divide-and-conquer strategy for semantic dependency parsing
نویسندگان
چکیده
We describe our SemEval2012 shared Task 5 system in this paper. The system includes three cascaded components: the tagging semantic role phrase, the identification of semantic role phrase, phrase and frame semantic dependency parsing. In this paper, semantic role phrase is tagged automatically based on rules, and takes Conditional Random Fields (CRFs) as the statistical identification model of semantic role phrase. A projective graphbased parser is used as our semantic dependency parser. Finally, we gain Labeled Attachment Score (LAS) of 61.84%, which ranked the first position. At present, we gain the LAS of 62.08%, which is 0.24% higher than that ranked the first position in the task 5. 1 System Architecture To solve the problem of low accuracy of long distance dependency parsing, this paper proposes a divide-and-conquer strategy for semantic dependency parsing. Firstly, Semantic Role (SR) phrase in a sentence are identified; next, SR phrase can be replaced by their head or SR of head. Therefore, the original sentence is divided into two kinds of parts, which can be parsed separately. The first kind is SR phrase parsing; the second kind is parsing the sentence in which the SR phrases are replaced by their head or SR of head. Finally, the paper takes graph-based parser as the semantic dependency parser for all parts. They are described in Section 2 and Section 4. Their experimental results are shown in Section5. Section 6 gives our conclusion and future work. 2 SR Phrase Tagging and Frame To identify SR phrase, SR phrase of train corpus are tagged. SR phrase is tagged automatically based on rules in this paper. A phrase of the sentence is called Semantic Role phrase (SR phrase) when the parent of only one word of this phrase is out of this phrase. The word with the parent out of the phrase is called Head of Phrase (HP). The shortest SR phrase is one word, while the longest SR phrase is a part of the sentence. In this paper, the new sequence in which phrases are replaced by their head or SR of head is defined as the frame. In this paper, firstly, SR phrases of the sentence are identified; secondly, the whole sentence is divided into SR phrases and frame; thirdly, SR phrase and frame semantic dependency are parsed; finally, the dependency parsing results of all components are combined into the dependency parsing result of the whole sentence. SR of HP is used as the type of this phrase. Only parts of types of SR phrases are tagged. In this paper, the tagged SR phrases are divided into two
منابع مشابه
A Divide-and-Conquer Strategy for Parsing
In this paper, we propose a novel strategy which is designed to enhance the accuracy of the parser by simplifying complex sentences before parsing. This approach involves the separate parsing of the constituent sub-sentences within a complex sentence. To achieve that, the divide-and-conquer strategy first disam-biguates the roles of the link words in the sentence and segments the sentence based...
متن کاملبرچسبزنی خودکار نقشهای معنایی در جملات فارسی به کمک درختهای وابستگی
Automatic identification of words with semantic roles (such as Agent, Patient, Source, etc.) in sentences and attaching correct semantic roles to them, may lead to improvement in many natural language processing tasks including information extraction, question answering, text summarization and machine translation. Semantic role labeling systems usually take advantage of syntactic parsing and th...
متن کاملMulti-stage Chinese Dependency Parsing Based on Dependency Direction
This paper presents a novel method for multistage dependency parsing based on dependency direction. In the method, dependency parsing processes are divided into multiple sub-stages, and each stage is in a sequential pattern, which makes it easier to take applicable solutions for different issues in dependency parsing. Meanwhile, dependency parsing in the previous stage provides a clearer contex...
متن کاملMultilingual Dependency Learning: A Huge Feature Engineering Method to Semantic Dependency Parsing
This paper describes our system about multilingual semantic dependency parsing (SRLonly) for our participation in the shared task of CoNLL-2009. We illustrate that semantic dependency parsing can be transformed into a word-pair classification problem and implemented as a single-stage machine learning system. For each input corpus, a large scale feature engineering is conducted to select the bes...
متن کاملAn improved joint model: POS tagging and dependency parsing
Dependency parsing is a way of syntactic parsing and a natural language that automatically analyzes the dependency structure of sentences, and the input for each sentence creates a dependency graph. Part-Of-Speech (POS) tagging is a prerequisite for dependency parsing. Generally, dependency parsers do the POS tagging task along with dependency parsing in a pipeline mode. Unfortunately, in pipel...
متن کامل