A neural system of phonematic transformation

• Autorzy: Podolak I. T. , Bielecki A.
• Języki publikacji: EN

Abstrakty

EN

A common task in speech processing for which neural networks are widely employed is text-to-phoneme conversion. In this paper we propose a novel solution to this problem by combining a multilayer neural network and a modular hybrid system that uses basic rules to subdivide the original problem into easier tasks which are then solved by dedicated neural networks. A hybrid solution can be more rapidly constructed than a single net solution, and is easily extendable. Input data representation is also discussed. A voting committee concept is used to enhance generalization abilities of the system. Efficiency of the proposed systems is compared.

Słowa kluczowe

EN

neural networks; input-output data representation; phonematic transformation

• Wydawca: Politechnika Gdańska
• Czasopismo: TASK Quarterly : scientific bulletin of Academic Computer Centre in Gdansk
• Rocznik: 2003
• Tom: Vol. 7, No 1
• Strony: 115--13018
• Opis fizyczny: 18 poz., rys., tab.

Twórcy

autor

Podolak I. T.

• Institute of Computer Science, Jagiellonian University, Nawojki 11, 30-072 Cracow, Poland

autor

Bielecki A.

• Institute of Computer Science, Jagiellonian University, Nawojki 11, 30-072 Cracow, Poland

Bibliografia

• [1] Digital Equipment Corporation, Maynard Mass,DECtalk DTC01 Owner’s Manual
• [2] Allen J 1985 From Text to Speech: the MITalk System, Cambridge University Press
• [3] Bolc L and Maksymienko M 1981 Text Processing Computer System, Warsaw UniversityPress, Poland (in Polish)
• [4] Sejnowski A and Rosenberg C R 1986 Technical Report86/01, John Hopkins University,Department of Electrical Engineering and Computer Science, USA
• [5] Bielecki A, Podolak I T, Wosiek J and Majkut E 1996 Acta Physica PolonicaB 27(9) 2253
• [6] Podolak I T, Lee S -W, Bielecki A and Majkut E 2000 Proc. Int. Conf. on Pattern Recognition,Barcelona, Spain, Vol. 1, pp. 720–724
• [7] Jakobson R and Halle M 1969 Language Foundations, PWN, Wroclaw (in Polish)
• [8] Steffen-Batogowa M 1975 Automatisation of Polish Language Phonematic Transcription,PWN, Warsaw (in Polish)
• [9] Karaś M and Madejowa M 1977 Polish Language Pronounciation Dictionary, PWN, Warsaw(in Polish)
• [10] Duch W and Jankowski N 1999 Neural Computing Surveys2163,http://www.icsi.berkeley.edu/ ̃jagota/NCS
• [11] Vapnik V 1998 Statistical Learning Theory, John Wiley and Sons
• [12] Baum E B and Haussler D 1989 Neural Computation1(1) 3
• [13] Podolak I T 1998 Problems of Data Representation in Layered Neural Networks, PhD Thesis,University of Mining and Metallurgy, Cracow, Poland (in Polish)
• [14] Podolak I T 1998 Comput. Phys. Commun.117(1–2) 62
• [15] Winston P H 1992 Artificial Intelligence, Addison Wesley
• [16] Haykin S 1994 Neural Networks: A Comprehensive Foundation, Maxwell Macmillan
• [17] McGarry K, Wermter K and McIntyre J 1999 Neural Computing Surveys262,http://www.icsi.berkeley.edu/ ̃jagota/NCS
• [18] Bax E 1998Neural Computation10(4) 975