MOST Deputy Minister Yu-Chin Hsu stated that he anticipates having the capability to accelerate the development of core technologies for Chinese voice dialogues. Regarding AI training, he explained that the most important component is accurate data. MOST established an AI voice digital database and invited Taiwan Tech to assist in the labeling of the voice data. Meanwhile, MOST has provided data that is convenient for everybody to access in order to accelerate the pace of AI training and improvement.

The release of this “AI Voice Data Set,” which contains 400 hours of recordings, will have the capacity to dramatically reduce the labor and time required for data labeling during the process of developing voice applications, and it will reduce the developmental barriers for related technologies.

At National Applied Research Laboratories (NARLabs), National Center for High-Performance Computing (NCHC), the NCHC DATA MARKET has been launched and is expected to attract more investment from enterprises and academic institutions in order to jointly promote a leap forward in the advancement Taiwan's voice AI technology and assist Taiwanese companies in undergoing digital transformations.

(TR/ Phil Sweeney)