A Semantic Composition Method for Deriving Sense Representations of Determinative-Measure Compounds in E-HowNet

A Semantic Composition Method for Deriving Sense Representations of Determinative-Measure Compounds in E-HowNet

本篇論文利用定量複合詞為例，示範如何利用廣義知網的語意合成機制來推導複合詞的語意及其表達式。首先我們定義了所有但有限數量的定詞跟量詞的廣義知網表達式，接著我們利用語意合成的規則針對任何新的定量詞去產生候選的語意表達式。然後我們在從調整語料集合去設計語意解歧規則，利用啟發式語意解歧規則跟參考上下文的詞來解決定量詞的廣義知網表達式的歧異，實驗顯示在語意推導跟解歧之後有88%的正確率。

In this paper, we take Determinative-Measure Compounds as an example to demonstrate how the E-HowNet semantic composition mechanism works in deriving the sense representations for all determinative-measure (DM) compounds which is an open set. We define the sense of a closed set of each individual determinative and measure word in E-HowNet representation exhaustively. We then make semantic composition rules to produce candidate sense representations for any newly coined DM. Then we review development set to design sense disambiguation rules. We use these heuristic disambiguation rules to determine the correct context-dependent sense of a DM and its E-HowNet representation. The experiment shows that the current model reaches 88% accuracy in DM identification and sense derivation.