|
1. |
A phonetic labeling method for mat database processing† |
|
Journal of the Chinese Institute of Engineers,
Volume 22,
Issue 5,
1999,
Page 529-534
Hsiao‐Chuan Wang,
Rong‐Liang Chiou,
Shiang‐Kai Chuang,
Yi‐Fen Huang,
Preview
|
PDF (641KB)
|
|
摘要:
This paper presents a semi‐automatic phonetic labeling method for processing in the MAT (Mandarin across Taiwan) speech database. MAT speech data are collected through the telephone networks. Each utterance has been transcribed into Chinese characters and Pinyin symbols. The proposed phonetic labeling method will mark the syllable and sub‐syllable boundaries in an utterance. Phonetic symbols are assigned to each segmented syllable. The segmentation process is accomplished by using hidden Markov modeling (HMM) and Viterbi decoding. The accuracy of syllable segmentation is detected by measuring the syllable length and the distance of a syllable from its state models. The experimental results show that the proposed labeling method can achieve segmentation accuracy around 90% for an allowed tolerance of 16 ms.
ISSN:0253-3839
DOI:10.1080/02533839.1999.9670491
出版商:Taylor & Francis Group
年代:1999
数据来源: Taylor
|
2. |
An overview of RNN‐based mandarin speech recognition approaches† |
|
Journal of the Chinese Institute of Engineers,
Volume 22,
Issue 5,
1999,
Page 535-547
Yuan‐Fu Liao,
Wei‐Tyng Hong,
Wern‐Jun Wang,
Yih‐Ru Wang,
Sin‐Horng Chen,
Preview
|
PDF (1416KB)
|
|
摘要:
The ANN‐based approach is an alternate to the HMM method which is the currently dominant technology in the field of speech recognition. This paper is an overview of the RNN‐based approaches to Mandarin speech recognition. Some RNN‐based approaches proposed previously: syllable‐boundary pre‐segmentation, broad‐class pre‐classification, prosodic phrase boundary detection, and isolated and continuous syllable recognitions are discussed. We find from the survey that these approaches are all effective. So the RNN technology is comparable to the conventional HMM method on Mandarin speech recognition.
ISSN:0253-3839
DOI:10.1080/02533839.1999.9670492
出版商:Taylor & Francis Group
年代:1999
数据来源: Taylor
|
3. |
Wavelet transforms for speech signal processing† |
|
Journal of the Chinese Institute of Engineers,
Volume 22,
Issue 5,
1999,
Page 549-560
Jhing‐Fa Wang,
Shi‐Huang Chen,
Jyh‐Shing Shyuu,
Preview
|
PDF (1183KB)
|
|
摘要:
The wavelet transform and its theory is one of the most exciting developments of the last decade. In fact, the wavelet transform has been developed independently for various different fields such as signal processing, image processing, audio and speech processing, communication, and mathematics. Due to the efficient time‐frequency localization and the multiresolution characteristics of the wavelet representations, the wavelet transforms are quite suitable for processing non‐stationary signals such as speech. In this paper, the wavelet transform and its theory will be first introduced, then comparisons between the wavelet transform and the classical short‐time Fourier transform approach to signal analysis will be provided. In addition, applying wavelet transforms in determining pitch, and segmenting consonant / vowel (C/V) parts as well as speech recognition will be discussed in this paper.
ISSN:0253-3839
DOI:10.1080/02533839.1999.9670493
出版商:Taylor & Francis Group
年代:1999
数据来源: Taylor
|
4. |
Lexical analysis for Chinese‐ difficulties and possible solutions† |
|
Journal of the Chinese Institute of Engineers,
Volume 22,
Issue 5,
1999,
Page 561-571
Keh‐Jiann Chen,
Preview
|
PDF (1181KB)
|
|
摘要:
Chinese sentences are composed with strings of characters without blanks to mark word boundaries. However, the basic processing unit for sentence processing is the word. It is the smallest meaningful, freely used unit for any natural language. Therefore lexical analysis became the first step in processing Chinese sentences. Usually a lexicon is utilized to match words and provide their syntactic and semantic information in the process of lexical analysis. During the word matching process, problems of segmentation ambiguity and occurrences of unknown words will occur. In this paper, both statistical methods and rule‐based methods are discussed for their advantages and disadvantages in solving segmentation ambiguities. For unknown word identification, off‐line word extraction methods and on‐line unknown word identification strategies are surveyed. Both methods complement each other in solving the problem. The strategies and knowledge sources for implementing a practical system are also discussed.
ISSN:0253-3839
DOI:10.1080/02533839.1999.9670494
出版商:Taylor & Francis Group
年代:1999
数据来源: Taylor
|
5. |
On phoneme‐to‐character conversion systems in Chinese processing† |
|
Journal of the Chinese Institute of Engineers,
Volume 22,
Issue 5,
1999,
Page 573-579
Wen‐Lian Hsu,
Yi‐Shiou Chen,
Preview
|
PDF (746KB)
|
|
摘要:
We propose a new goal for constructing a Chinese phoneme‐to‐character automatic conversion system. Instead of trying to get the largest number of correct characters converted, this goal strives for extracting the correct events from a given sentence. We believe this objective is more fruitful in that the extracted events can be better utilized in other applications such as information retrieval, extraction and dialogue systems. Furthermore, it would motivate researchers to work on the more challenging problem of natural language understanding. An information map is proposed for knowledge representation. These maps give very concise summary of different aspects of an object. They also provide more convenient ways to implement semantic template matching.
ISSN:0253-3839
DOI:10.1080/02533839.1999.9670495
出版商:Taylor & Francis Group
年代:1999
数据来源: Taylor
|
6. |
Towards a pure Chinese spoken dialogue system† |
|
Journal of the Chinese Institute of Engineers,
Volume 22,
Issue 5,
1999,
Page 581-591
Tung‐Hui Chiang,
Yi‐Chung Lin,
Preview
|
PDF (1182KB)
|
|
摘要:
This paper gives an overview of our spoken dialogue system, which is designed to meet four objectives, Portability, Usability, Robustness, and Extensibility (PURE). First, an integrated scoring function is proposed so that all components of the spoken dialogue system can be properly integrated, and information from all components can be considered jointly to resolve ambiguities. Moreover, instead of shallow analysis or simply keyword/key‐phrase matching, deeper analysis is conducted by incorporating robust parsing, error recovery and semantic analysis to improve discrimination and robustness of the language understanding component. To achieve PURE design of dialogue management, a domain transparent framework is also proposed in this paper. Results and experiences we got from building systems of various domains showed very encouraging results.
ISSN:0253-3839
DOI:10.1080/02533839.1999.9670496
出版商:Taylor & Francis Group
年代:1999
数据来源: Taylor
|
7. |
Executable EFSM‐based data flow and control flow protocol test sequence generation using reachability analysis |
|
Journal of the Chinese Institute of Engineers,
Volume 22,
Issue 5,
1999,
Page 593-615
Chung‐Ming Huang,
Ming‐Yuhe Jang,
Yuan‐Chuen Lin,
Preview
|
PDF (2070KB)
|
|
摘要:
This paper presents an executable Extended‐Finite‐State‐Machine (EFSM)‐based protocol test sequence generation method with specified variable bound. The proposed method overlaps data flow and control flow tests. In the data flow part, the transition paths that containdefinitionuses andoutputuses of variables in the protocol specifications need to be detected and tested. An executable data flow test path, which is called an EDO‐path, contains two parts: (1) a switching transition sequence that directs to a transition T containing adefinitionuse of a variable V, and (2) a transition sequence originated from T, such that its head transition has adefinitionuse of V and its tail transition contains either (i) an output use of V, or (ii) an output use of another variable which is affected by this definition use of V. In the control flow part, each transition must be tested as to whether it conforms to the original specification or not. An executable control flow test path, which is called an EC‐path, for a transition T contains (1) a switching sequence that directs to the head state of T, (2) the transition T itself, and (3) the Unique Input/Output sequence of T's tail state, which verifies whether the tail state of T is correct or not. To shorten the length of the total test sequences, the proposed method overlaps the control flow test sequences with the data flow test sequences. That is, a test sequence may combine both control flow and data flow tests. The overlap is achieved as follows: when of searching EDO‐paths, EC‐paths are also searched in the same executable transition exploration sequences. If not all of the EC‐paths are overlapped with EDO‐paths, the remaining EC‐paths are searched individually. Additionally, to provide flexible protocol conformance testing, protocol testers can specify some variable bound for their specific test purposes. That is, protocol testers can run tests they are interested in by specifying the associated variable bound in the protocol specification.
ISSN:0253-3839
DOI:10.1080/02533839.1999.9670497
出版商:Taylor & Francis Group
年代:1999
数据来源: Taylor
|
8. |
Quasi time‐optimal PID control of multivariable systems: A seesaw example |
|
Journal of the Chinese Institute of Engineers,
Volume 22,
Issue 5,
1999,
Page 617-625
Chia‐Ju Wu,
Preview
|
PDF (771KB)
|
|
摘要:
A numerical approach is proposed to design PID controllers for multivariable (multi‐input or multi‐output) systems. Different from traditional tuning methods, the PID parameters in the proposed method are determined by solving a nonlinear programming (NLP) problem. Since the objective function of the NLP problem is the traveling time between two states, the PID controller will drive the system from a given initial state to the desired final state in a quasi time‐optimal manner. The reason why the PID controller cannot generate truly time‐optimal movement will also be given. A seesaw system that has one input and two outputs will be used for simulation and practical implementation to illustrate the validity of the proposed method.
ISSN:0253-3839
DOI:10.1080/02533839.1999.9670498
出版商:Taylor & Francis Group
年代:1999
数据来源: Taylor
|
9. |
A self‐tuning run‐by‐run process controller for processes subject to random disturbances |
|
Journal of the Chinese Institute of Engineers,
Volume 22,
Issue 5,
1999,
Page 627-638
Ruey‐Shan Guo,
Jin‐Jung Chen,
Argon Chen,
Shui‐Shong Lu,
Preview
|
PDF (942KB)
|
|
摘要:
In this paper a self‐tuning run‐by‐run process controller is presented. The controller has the capability of choosing a control parameter dynamically in response to the underlying process disturbances. There are two modules in this controller: a self‐tuning loop trigger module and a run‐by‐run feedback control module. In the self‐tuning loop trigger module, two EWMA control charts are used sequentially to determine if there is a large or medium shift in the process output and to trigger a new self‐tuning loop accordingly. In the run‐by‐run feedback control module, the control parameter and control model are re‐tuned sequentially and a new process recipe is generated, on a run‐by‐run basis, to compensate for the process output's deviation from the target. Monte Carlo simulation results show that the self‐tuning run‐by‐run process controller is superior to the current run‐by‐run process controller with a fixed control parameter.
ISSN:0253-3839
DOI:10.1080/02533839.1999.9670499
出版商:Taylor & Francis Group
年代:1999
数据来源: Taylor
|
10. |
A numerical study on turbulent vortex‐shedding flows around a cubical form |
|
Journal of the Chinese Institute of Engineers,
Volume 22,
Issue 5,
1999,
Page 639-648
Yih‐Ferng Peng,
RobertR. Hwang,
Preview
|
PDF (1001KB)
|
|
摘要:
The turbulent vortex shedding flow behind a square cube is numerically analyzed by the LES approach. The eddy viscosity and Smagorinsky models have been used to simulate the SGS Reynolds stress and eddy viscosity, respectively. The performance of the computations is evaluated by comparing with available measurements, as well as with those results obtained from RANs. It is found that the LES approach predicts the Strouhal number and the mean stream‐wise velocity profiles with a better accuracy than RANs.
ISSN:0253-3839
DOI:10.1080/02533839.1999.9670500
出版商:Taylor & Francis Group
年代:1999
数据来源: Taylor
|
|