NCSoft Co., Ltd. (hereinafter referred to as NC) will present three papers at the international artificial intelligence (AI) conference ‘INTERSPEECH 2022’.
InterSpeech, now in its 23rd year, is the world’s most prestigious speech AI conference where global AI companies share the latest technologies. InterSpeech 2022 will be held from the 18th to the 22nd of this month at Convensia in Songdo, Incheon.
The ‘Speech AI Lab’ under the NC AI Center has been conducting research on speech synthesis for a long time and has shared the results with prestigious academic societies. Papers have been published on InterSpeech for three consecutive years, and three papers were accepted this year.
The topics of the three papers are ▲study on a method of separately modeling timbre and pitch expression based on adversarial multi-task learning ▲study on improving speech quality using a pre-trained neural vocoder ▲research on improving naturalness by diversifying the expressiveness of synthesized sounds. .
Through this research, the quality and naturalness of voice synthesis (TTS) and vocal synthesis (SVS) systems were realized to the highest level in the industry.
Kim Hee-man, director of NC Speech AI Lab, said, “NC’s voice AI technology has already gone beyond commercialization and is looking toward the next step,” adding, “We plan to research differentiated technology and continuously share the results with the outside world.” He said.
Meanwhile, NC will set up a booth at the InterSpeech conference to share research status and demonstrate a demo version. At the booth, conference participants can check the level of NC’s voice AI technology and also receive consultation on hiring AI personnel.
Source: Pangyo Techno Valley Official Newsroom
→ Go to ‘Asian Innovation Hub Pangyo Techno Valley 2022’ news