Louis Wei-lun Lu
In this synchronic study, I shall adopt a corpus-based approach to investigate the semantic change of V-diao in Mandarin. Semantically, V-diao constructions fall into three categories:
A) Physical disappearance from its original position, with the V slot filled by physical verbs, such as tao-diao ¡§escape,¡¨ diu-diao ¡§throw away,¡¨ and so on.
B) Disappearance from a certain conceptual domain, rather than from the physical space, with the V slot filled by less physically perceivable verbs, such as jie-diao ¡§quit,¡¨ wang-diao ¡§forget,¡¨ and the like.
C) The third category of V-diao involves the speaker¡¦s subjective, always negative, attitude toward the result. Examples include: lan-diao ¡§rot,¡¨ ruan-diao ¡§soften,¡¨ huang-diao ¡§yellow,¡¨ and so forth.
It is claimed in this paper that the polysemy between types A and B is motivated by metaphorical transfer [Sweetser, 1990; Bybee, Perkins and Pagliuca, 1994; Heine, Claudi and Hunnemeyer, 1991]. Based roughly on Huang and Chang , I demonstrate that a cognitive restriction on selection of the verb will cause further repetitive occurrence of negative verbs in the V slot. Finally, I shall claim that pragmatic strengthening [Hopper and Traugott, 1993; Bybee, Perkins and Pagliuca, 1994] contributes to the emergence of unfavourable meaning in Type C.
Hopefully, this research can serve as a valid argument for the interaction of language use and grammar, and the conceptual basis of human language.
metaphorical transfer, pragmatic strengthening, conceptualization
Chuan-Jie Lin, Che-Chia Liu, Hsin-Hsi Chen
Captions in videos contain valuable information for video retrieval. Although texts in captions can be obtained easily in the new image compression formats like MPEG2, there still are many video programs encoded in older formats. Thus, video OCR is indispensable for content-based video retrieval. This paper proposes a simple video OCR method for Chinese captions, including image capturing, caption region deciding, background removing, character segmentation, OCR and post-processing. We employed Discovery Channel films as training and testing corpus. In an outside test, the accuracy of the video OCR was 84.1%. The hardware used in the experiment consisted of a computer with a P4-1.7G CPU, 256MB RAM and a 40G, 7200rpm hard disk. On average, it took 29 minutes and 11 seconds to process a film 495MB in size. We also applied the results of video OCR to video retrieval and question answering.
digital library, question answering, Chinese video OCR, video retrieval
Jau-Hung Chen, Yung-An Kao
In a text-to-speech (TTS) conversion system based on the time-domain pitch-synchronous overlap-add (TD-PSOLA) method, accurate estimation of pitch periods and pitch marks is necessary for pitch modification to assure optimal quality of synthetic speech. In general, there are two major tasks in pitch marking: pitch detection and location determination. In this paper, an adaptable filter, which serves as a bandpass filter, is proposed for use in pitch detection to transform voiced speech into a sine-like wave. The pass band of the adaptable filter can be adapted based on the fundamental frequency. Based on the sine-like wave, a peak-valley decision method is proposed to determine the appropriate parts (positive part and negative part) of voiced speech for use in pitch mark estimation. In each pitch period, two possible peaks/valleys are searched, and dynamic programming is performed to obtain pitch marks. Experimental results indicate that our proposed method performs very well if correct pitch information is estimated.
Jason S Chang, David Yu, Chun-Jun Lee
Machine Translation is one of the most difficult problems in the field of natural language processing. In the past, MT has been applied to professional communication in the process of translating technical and corporate document on a specific domain. Recent years saw the rapid development of Internet as a new form of communication and information exchange, and the need to access information across the language barrier became apparent. People began to look into the role that MT can play in Cross Language Information Retrieval. The prevalent approach to CLIR is based on translation of query, in particular query phrases. However, for CLIR there is an additional new objective of translating into something that is relevant to the collection being searched upon. Therefore, the current approach of using general bilingual word list or an off-the-shelf commercial MT software is bound to be very ineffective in terms of retrieving relevant documents. We propose a new approach to Statistical Phrase Translation Model (SPTM), aimed at achieving a tighter estimation of phrase translation. Experiments were conducted using bilingual phrases in BDC Electronic Chinese-English Dictionary. Preliminary results shows the approach is much faster and produces better word alignment for phrases, which has not been possible using previous approaches.
Statistical Machine Translation; Phrase Translation; Cross-language Information Retrieval