(From Miao Jianming of ThinkIT Lab) The HNC language processing team wins the first prize on the Chinese event description unit identification in the syntax analysis evaluation and wins second prize on the Chinese function block analysis. The evaluation is undertaken by the Tsinghua University, the Northeast University and sponsored by Chinese Information Processing Society of China. There are 24 teams, coming from America, Europe, mainland of China, Hong Kong and the Taiwan area, to participate the evaluation.
Effective and authentic evaluation on syntax analysis is an important way to promote natural language processing. CoNLL (Conference on Computational Natural Language Learning) in English language is a world-famous conference on natural language processing, whose mission includes identification of basic noun phrase, analysis of text block, identification of clause, identification of denomination object, mark of meaning part, dependence analysis, syntax analysis and integral processing of meaning part. The evaluation also covers some sub-projects on natural language processing. On Chinese language, whether the syntax analysis based on block analysis is good or bad determines effectiveness of syntax analysis system. The evaluation (CIPS-ParsEval-2009), sponsored by Chinese information Processing Society of China, is the top-class competition on the information processing of syntax semantic. There are five assignments in the evaluation, which are Chinese basic block analysis, Chinese function block analysis, unit identification of event description and identification of syntax construction tree. The goal is to completely evaluate the technology of syntax analysis in various lays for Chinese 'character/phrase-block-sentence'. The processing software is expected to be developed via the evaluation, in order to fulfill various applications in Chinese syntax analysis in various lays.
Chinese block is combination of syntax semantic information and vocable conjunction in the block bridges the conjunction of syntax semantic. When the syntax analysis system can offer explicit syntax criterion and generate rational semantic explanation. The HNC language processing team possesses unique theoretic basis and accumulation in core technology in natural language processing. Since 1998, they began to work over the core technology of natural language, which centered the Chinese semantic block and sentence species. During the Chinese language research, HNC processing technology introduces the technology of dependence tree to analyze sentences, differing from the common ways in modern Chinese. With a few years effort, the team has made some breakthrough and put forward their own opinion in the projects related to natural language processing, especially on the analysis of function block and frame of event distribution. All of the above have laid foundations for their achievements in the competition. In the near future the team will dedicate themselves to engineering application of the technology.