Sixth International Conference on Spoken Language Processing
October 16-20, 2000
Data-Driven Importance Analysis of Linguistic and Phonetic Information
Achim F. Müller (1,3), Jianhua Tao (1,2), Rüdiger Hoffmann (3)
(1) Siemens Corporate Technology, Munich, Germany
In this paper the weight decay concept known from neural
network theory is applied to the two modules involved
in prosody generation within our text-to-speech system
Papageno. Both modules are based on neural networks (NN).
Preprocessing layers are inserted connected to the inputs
of specialized NN architectures via diagonal weight matrices.
The weight decay concept is applied to the weights of
these diagonal weight matrices. This allows an importance
analysis of the used input parameters in the context of the
used NN architectures.
(2) Tsinghua University, Beijing, China
(3) Dresden University of Technology, Germany
In the symbolic prosody module the importance for phrase
break prediction of part-of-speech (POS) tags could be
evaluated Further, the necessary length of
a POS context window could be analyzed and optimized.
For fO-generation for Mandarin language an importance
analysis of the phonological information could be
performed. The importance analysis led to an optimized
input feature set reducing the squared error of the used NN
Müller, Achim F. / Tao, Jianhua / Hoffmann, Rüdiger (2000):
"Data-driven importance analysis of linguistic and phonetic information",
In ICSLP-2000, vol.2, 75-78.