Frank Kao-PingK Soong

Title: Frank Kao-PingK Soong: Innovator in Bilingual TTS Techniques

Introduction

Frank Kao-PingK Soong is a prominent inventor based in Beijing, China. He has made significant contributions to the field of text-to-speech (TTS) technology, particularly in bilingual applications. With a total of 3 patents to his name, Soong's work has advanced the capabilities of multilingual speech generation.

Latest Patents

One of Soong's latest patents is focused on HMM-based bilingual (Mandarin-English) TTS techniques. This exemplary method generates speech based on text in one or more languages. It includes providing a phone set for two or more languages and training multilingual hidden Markov models (HMMs) that share states across languages. The process involves receiving text in one or more of the languages and generating speech based on the multilingual HMMs. Other techniques in this patent include mapping between decision trees for different languages and Kullback-Leibler divergence analysis for multilingual TTS systems.

Another notable patent by Soong is related to constrained line search optimization for discriminative training of HMMs. This method optimizes a continuous density hidden Markov model (CDHMM) by imposing constraints for discriminative training. It approximates an objective function as a smooth function of CDHMM parameters and performs a constrained line search to optimize these parameters. Various other methods, devices, and systems are also disclosed in this patent.

Career Highlights

Frank Soong is currently associated with Microsoft Technology Licensing, LLC, where he continues to innovate in the field of speech technology. His work has been instrumental in enhancing the performance and accuracy of TTS systems, particularly for bilingual applications.

Collaborations

Soong has collaborated with notable colleagues, including Peng Liu and Yi-Jian Wu. Their combined expertise has contributed to the advancement of TTS technologies and the development of innovative solutions in the field.

Conclusion

Frank Kao-PingK Soong is a key figure in the development of bilingual TTS technologies, with several patents that showcase his innovative approaches. His contributions continue to shape the future of speech generation systems, making them more accessible and effective for multilingual users.