*To implement, improve and handle CSR language modeling and automatic pronunciation generation for multiple languages
* To improve voice recognition accuracy and out-of-vocabulary rejection rate for noisy speech, kid’s speech, and emotional speech with limited CPU and memory for multi languages
* Improve runtime voice recognition and sample voice applications for PlayStation product for many languages
* Develop robust automatic speech recognition (ASR) technologies to various kinds of distortions and variations such as channel and environment distortions, emotional speech, variety of speaking rate and speaking style for multiple languages
* Integrate ASR with technologies of microphone array, echo cancellation, speech enhancement, and other DSP processing for far-field voice recognition under noise
* Implement and improve very large vocabulary continuous voice recognition performance under variety of game sounds and music background
* Improve acoustic model training and adaptation
Reply With Quote