Development of the philips 1999 taiwan Mandarin benchmark system
نویسندگان
چکیده
This paper describes the Philips Large Vocabulary Continuous Mandarin speech recognition system for the 1999 Taiwan benchmark. The basic system architecture is based on the Philips LVCSR technology developed for Western languages. However, several modifications are made in order to better suitted processing Chinese spoken languages. In the paper, we present some experimental results on the two tasks we participated in this benchmark: digit and continuous syllable. For the development set, we were able to obtain a digit/syllable error rate of 2.9%/23.9%. At the final evaluation, our system achieves the lowest error rate of 3.1%/24.3% among all participating sites.
منابع مشابه
Improvements of the Philips 2000 Taiwan Mandarin benchmark system
In this paper, we present the Philips large vocabulary continuous Mandarin speech recognition system developed for the 2000 Taiwan Speech Input Technology Assessment. We systematically integrated key Mandarin components with up-todate Western-language techniques to build up a state-of-the-art Mandarin speech recognition system. These technologies include robust pitch extraction/tone modeling, c...
متن کاملMAT-2000 - design, collection, and validation of a Mandarin 2000-speaker telephone speech database
Mandarin speech data Across Taiwan (MAT) is a project initiated by members of the Association for Computational Linguistics and Chinese Language Processing (ACLCLP) to collect speech data through public telephone networks in Taiwan. Totally over 7000 Taiwanese individuals have provided speech data. The results were released as a series of MAT speech databases to the research community in Taiwan...
متن کاملBiopharmaceutical Innovation System and the Influence of Policies: The Case of Taiwan (2000-2008)
This article discusses the influence of policies on the development of biopharmaceuticals. We choose the experiences of Taiwan for our empirical study and focus on the evolution between 2000 and 2008; in the period of time the country provides an interesting example for further exploration of biopharmaceutical policies. Among all the policies, the two National Programs (National Research Progra...
متن کاملPhonetic Modelling in the Philips Chinese Continuous Speech Recognition System
We have extended the Philips large vocabulary continuous speech recognition system towards Chinese On the way from our existing Western language technology to Mandarin the rst step was to build a suitable phonetic model This paper describes the development of our phonetic model excluding tones for Mandarin Chinese We will present a systematic comparison of three forms of sub syllabic units for ...
متن کاملOn the Argument Structures of the Transitive Verb 'annoy; be annoyed; bother to do': A Study Based on Two Comparable Corpora
This paper investigates the transitive uses of the verb fan „annoy; be annoyed; bother to do‟, which exhibit both similarities and disparities between Beijing Mandarin and Taiwan Mandarin, as far as the data from Gigaword corpus, containing data from Mainland China (XIN) and Taiwan (CNA), are concerned. In terms of similarities, the causative (and agentive) use(s) of the transitive fan is/are s...
متن کامل