Sixth International Conference on Spoken Language Processing
In this paper, we propose a method for domain unconstrained language understanding based on the linguistic toolsets CKIP-AutoTag and How-net, and the devised Acting Role Table (ART).
In our approach, the analysis of article is performed sentence by sentence. For each sentence in the article, word segmentation is first performed using CKIP-AutoTag. Next, the semantics of each sentence is represented in the devised Acting-Role-Table (ART). The ART consists of acting roles of the sentences, i.e., action, agent, instrument, theme, location, and time, together with their associated modifiers. In this step, we use verb-driven syntax analysis to determine the acting roles in the sentences; and use semantic analysis to constrain time acting roles based on time features defined in How-net. Finally, the semantic network is constructed to record the relationship of each sentence.
To test whether our approach for text understanding is feasible, an auto reading comprehension system is built trying to answer the exercises of the primary school textbook. The exercises contain many questions related to the article. Seven textbooks ranged from third-grade to ninth-grade are used as our testing articles. Most of the questions in the exercises can be answered by our system if the answers can be derived from the article.
Bibliographic reference. Wang, Jhing-Fa / Wang, Hsien-Chang / Lee, Kin-Nan / Huang, Chieh-Yi (2000): "Domain-unconstrained language understanding based on CKIP-auto tag, how-net, and ART", In ICSLP-2000, vol.3, 478-481.