No description available.

Generating computer animation from natural language instructions is a complex task that encompasses several key aspects of artificial intelligence including natural language understanding, computer graphics and knowledge representation. Traditionally, this task has been approached using rule based systems which were highly successful on their respective domains, but were difficult to generalise to other domains. In this thesis, I describe the key theories and principles behind a domain-independent machine learning framework for constructing natural language based animation systems, and show how this framework can be more flexible and more powerful than the prevalent rule based approach. / I begin this thesis with a thorough introduction to the goals of the research. I then review the most relevant literature to put this research into perspective. After the literature review, I provide brief descriptions to the most relevant technologies in both natural language processing and computer graphics. I then report original research in semantic role labelling and verb sense disambiguation, followed by a detailed description and analysis of the machine learning framework for natural language based animation generation. / The key contributions of this thesis are: a novel method for performing semantic role labelling of prepositional phrases, a novel method for performing verb sense disambiguation, and a novel machine learning framework for grounding linguistic information in virtual worlds and converting verb-semantic information to computer graphics commands to create computer animation.

本研究探討漢語動補結構在語料庫中的使用情況，透過語料庫研究方法，辨析動詞「用X」及「弄X」與動補結構的關係，並探究動補結構中「用X」及「弄X」相互替代的使用現象，最後由詞彙語意及語義韻律的角度提出說明。 本研究利用中文詞彙網路（Chinese Wordnet）、《平衡語料庫》（Sinica Corpus 4.0）以及《批踢踢語料庫》（PTT Corpus），觀察語料庫中動詞「用」、動詞「弄」、動補結構「用X」及動補結構「弄X」的語料分布，以及語料與動詞義項的互動情況。研究發現，在語料庫中，動詞「用」的主要語義是「利用特定對象的特定功能」，而動詞「弄」的主要語義則是以「造成特定結果」和「代動詞，做」，且在語料庫中，「用X」結構的分布集中、「弄X」分布廣泛。研究亦發現，《平衡語料庫》中的動補結構語料相對來說較《批踢踢語料庫》中的語料組合穩定，此一現象是由於不同語料庫中的語料結合緊密程度不同所致。 藉由動詞代換檢測及動詞義項的比對，可知在動結式動補結構這個特定的環境下，動詞「用」和「弄」顯示了與近義詞相似的特質，甚至展現出詞彙語意關係中的上下位（hyponymy）或者蘊含關係（entailment）。研究最後透過觀察「用X」和「弄X」動補結構的語義韻律，探究動詞「用」和「弄」在動詞義項以外的異同之處，並發現語料庫中的動補結構「用X」及「弄X」在語義韻律的分布上有相似的表現。 / This paper aims to explore the cause of the mutual substitution of verbs yòng(用) and nòng(弄) in verb-complement structure in Taiwan Mandarin. The study adopts a semantic corpus-based approach to examine this phenomenon. To find out the resemblance of meanings shared by the verbs yòng and nòng, this study first retrieved their senses by using Chinese Wordnet and analyzed the sentences from the Sinica Corpus. Three hundred instances of yòng and 240 instances for nòng were analyzed and classified according to their corresponding verb senses. The results showed that yòng and nòng both contain the concept of “utilize certain tool/method/means to fulfill particular purposes”, proving that they could possibly form a synonymous set of verb. The analysis then focused on the behavior of the verbs in verb-complement structures. The verb-complement structures of yòng and nòng from the Sinica Corpus and the PTT Corpus were further inspected: 14 entries of ‘yòng-complement’ (179 instances) and 32 entries of ‘nòng-complement’ (157 instances) were found in Sinica Corpus, while 21 entries of ‘yòng-complement’ (82 instances) and 97 entries of ‘nòng-complement’ (781 instances) were found in the PTT Corpus. The distributions showed that the verb-complement structures in Sinica Corpus were relatively more stable than those in the PTT Corpus, and the phenomenon was inferred to be caused by the distinct degree of the structure combination. In order to examine the possibility of the mutual substitution of yòng and nòng, a verb-substitution test was conducted. The results were further examined by comparing with the verb-complement structures found in the PTT Corpus. In addition, introspective survey of native speakers of Mandarin was also added to re-confirm the substitution test result. Lastly, the verb senses and the semantic prosody of the verb-complement structures were also analyzed. The study showed that the verbs yòng and nòng have shared senses when they appear in resultive verb-complement structures, and the mutual substitution patterns were also discovered. As for the analysis of semantic prosody of the verb-complement structures, the result showed that yòng-complement and nòng-complement have identical distribution tendency: structures with neutral semantic prosody occupied the greatest proportion, followed by the negative ones and the positive ones, which also proved the synonymous relationship between the verbs yòng and nòng in the verb-complement structures.