China Smart Voice Market Size Grows 44% in 2021
|A volunteer helps a visually impaired man to read with a text-to-speech device at a library in Wuyi County, Jinhua City, east China’s Zhejiang Province, Oct 13, 2021. [People’s Daily Online/ Zhu Hui]|
A virtual host named “i+” has been put into use for the upcoming Beijing 2022 Olympic and Paralympic Winter Games. Equipped with intelligent voice recognition technology, it can translate Mandarin Chinese into English, French, Japanese and other languages. other languages in real time, spreading Beijing 2022 knowledge to the world faster.
Intelligent voice recognition technology is increasingly integrated into terminal applications to serve people’s daily lives. Besides the AI host, its application scenarios also include a smart fitness mirror, smart wearable devices that can help couriers collect and deliver parcels, and a smart mouse that can tap automatically when you speak.
The intelligent voice industry, as an important part of the software industry, has entered a new stage characterized by high-speed development, said Wang Jianwei, deputy director of the information technology department at the Chinese Ministry of Industry and Information Technology (MIIT), during the China Intelligent Voice Industry Development Summit Forum held in December 2021.
China’s smart voice industry has seen dynamic growth in recent years, with breakthroughs in core technologies, Wang said, adding that the voice recognition accuracy rate has reached 98%.
China’s smart voice market size reached 21.7 billion yuan ($3.4 billion) in 2020, an increase of 31% year-on-year, and is expected to grow 44% year-on-year to reach 28.5 billion yuan in 2021, effectively boosting industrial digitalization, pointed out a white paper on the development of China’s smart voice industry (2020-2021) released on December 18, 2021.
In a world where everything is connected, more smart devices need to be controlled remotely, which creates opportunities for the industry, said Liu Qingfeng, chairman of the board of the Speech Industry Alliance of China (SIAC) and president of iFlytek, a leading Chinese AI company.
The number of smart devices driven by the growing demand for voice interaction is growing rapidly, Liu noted, adding that the amount of interaction services handled by the company’s voice assistants saw an 84 percent increase in 2021 year over year.
|Students explore the moon using virtual reality (VR) headsets at a primary school in Qingdao city, east China’s Shandong province, Sept 18, 2021. [People’s Daily Online/ Wang Haibin]|
Intelligent voice technology faces three major challenges, including multilingual intercommunication, human-computer interaction in complex scenarios and multimodal virtual world, Liu said.
Multilingual intercommunication includes not only foreign languages, but also dialects in China; effective interaction in complex scenarios requires accurate voice recognition when multiple people are talking at the same time; and multimodal interaction means adding timbre, pitch, expression, mouth shape and other factors to the voice to make speech recognition smarter, Liu said.
The recognition rate of iFlytek products is estimated to have increased from 69% to 80% under complex scenarios in 2021, according to Liu.
Key innovation drivers for the future development of the smart voice industry include unsupervised learning, multimodal fusion, and innovative interdisciplinary brain science research, the white paper says.
The algorithm regarding unsupervised learning and the low-resource model algorithm still need major breakthroughs; and in terms of AI chips, the basis of computing power, China must hurry to catch up with world leaders.
According to Wang, MIIT plans to continue the high-quality development of the intelligent voice industry from three aspects.
|A robot helps popularize law knowledge among students in Yuanhui District, Luohe City, central China’s Henan Province, Dec 13, 2021. [People’s Daily Online/ You Yahui]|
The ministry will call on regional departments to speed up the formulation of industrial policies that promote the integrated development of intelligent voice technology and the real economy, he noted.
In the meantime, he intends to encourage leading companies and scientific research institutions to join forces for new technological breakthroughs in order to continuously improve technologies related to speech recognition, synthesis, interaction and voice chips, and build national public service platforms for intelligent voice testing to support the development of the industry, the official said.
SIAC has already attracted more than 70 enterprises with core technologies along the industrial chain. At least 70 more will join the alliance in the future, with more research institutes and universities expected, according to Wang.
MIIT will also expand the application scenarios of voice technology so that it can be integrated into smart manufacturing, smart home, smart health, education, elderly care and other fields.
(Source: People’s Daily Online)