Collect
Check out my collection
0
Useful +1
"AI synthetic anchor" is the world's first fully simulated intelligent AI host jointly released by Sogou and Xinhua News Agency at the Fifth World Internet Conference on November 7, 2018.
[1-2]
Through speech synthesis, lip synthesis, expression synthesis and deep learning technologies, an "AI synthetic anchor" with the same broadcasting ability as a human anchor has been cloned.
[3]
In May 2020, the world's first 3D AI synthetic anchor"
New Xiaowei
At...
National two sessions
Officially unveiled on the eve of the opening.
[4]
- Chinese name
- AI synthetic anchor
- Release meeting
- The fifth World Internet Conference
- Release time
- November 7, 2018
- nature
- AI host
- Issuing authority
- Sogou and Xinhua News Agency
- Adopt technology
- Speech synthesis, lip synthesis, expression synthesis, etc
The "AI synthetic anchor" is supported by Sogou's "split body" technology.
[5]
Through a number of cutting-edge technologies such as face key point detection, face feature extraction, face reconstruction, lip recognition, and emotion transfer,
[5]
And combined with multi-modal information such as speech, image and other joint modeling training, to generate the AI avatar model indistinguishable from the real person.
[6]
The technology can automatically generate a video of the corresponding content from the input Chinese and English text, and ensure that the audio and facial expressions and lip movements in the video remain natural and consistent, showing the same information transmission effect as a human anchor.
"AI synthetic anchor" not only has the Chinese "AI synthetic anchor", but also the English "AI synthetic anchor" with the mission of "connecting China and the world and communicating with the world".
[2]
Establish the association between the input text and the output audio information. In the image generation engine, the industry's Sogou face recognition, three-dimensional face reconstruction, expression modeling and other technologies are used to conduct feature learning and modeling of facial expression actions, establish the association mapping between the input text, output audio and output visual information, and finally generate the output dope-body video.
[7]
On November 7, 2018, the fifth World Internet Conference kicked off in Wuzhen, Zhejiang Province. Wang Xiaochuan, CEO of Sogou Company, Liu Siyang, vice president of Xinhua News Agency, Wang Yanfeng, general manager of intelligent Voice Division of Sogou Company, and Qiu Hao, host, jointly released AI synthetic anchors.
[1]
On February 19, 2019, Sogou Company and
Xinhua News Agency New Media Center
Jointly released the first standing AI synthetic anchor, and the new AI synthetic anchor upgraded from the past "sitting news broadcast" to the "standing broadcast" combined with body movements, marking another breakthrough in "Sogou split body" technology.
[8]
In March 2019, Xinhua News Agency and Sogou Company recently released a newly upgraded standing AI synthetic anchor in Beijing, and launched the world's first AI synthetic female anchor.
[9]
On May 21, 2020, Sogou and Xinhua News Agency launched the 3D AI synthetic anchor"
New Xiaowei
"
[4]
[10]
. Compared with the previous versions of "New Xiaohao" and "New Xiaomeng", the AI synthetic anchor named "New Xiaowei" is highly responsive to real people's hair and skin, and has made a big jump in many aspects such as three-dimensional sense, flexibility, plasticity, and interaction ability, and can also make a variety of postures and expressions closer to real people according to the content of the broadcast.
[11]
Russian media AI weather forecaster [17]
On April 11, 2023, Kuwait's first artificial intelligence (ai) generated female anchor debuted, attracting a number of media attention and reports around the world, and also caused a great response on social media.
[18]
In June 2023, Radio Television Hong Kong's AI weather anchor Aida officially debuted
,
Starting from 26 June, the local and international weather conditions will be reported in the Radio Television Hong Kong Weather Brief programme.
[19]
In August 2023, Shandong Radio and Television Station meta-universe and Artificial Intelligence Laboratory amplified the recruitment, and the country's first super realistic male digital anchor "Dai Qing" was officially unveiled
.
[20]
The "AI synthetic anchor" uses the real image of the Chinese and English anchors of Xinhua News Agency, and simulates the real person broadcast picture with the voice and synthesis technology of Sogou's "dopes". This form of broadcasting breaks through the previous speech and image synthesis field, which can only create virtual images, and cooperate with the lip effect of speech output, and improve the truth degree of audience information acquisition. Using the "Sogou split body" technology, the "AI virtual anchor" can also effectively output audio and video synthesis effects in real time, and users will get real-time broadcast video after entering text through various ways such as text typing, voice input, and machine translation. This mode of operation will reduce the costs of news media in post-production and improve the production efficiency of news videos.
[12]
The "AI synthetic anchor" has the same broadcasting ability as a human anchor, and can broadcast 24 hours a day.
[12]
The new version of AI synth anchor uses wavernn waveform modeling technology to replace the traditional vocoder and achieve audio synthesis. Through model optimization and the use of multi-style data, the new version of AI synthetic anchor realizes expression synthesis, body movements and semantic matching.
[13-14]
With the optimization of the graphics generation engine, the AI synthetic anchor has been upgraded from the past "sitting news broadcast" to the "standing broadcast" combined with body movements, and the use of "Sogou split body" technology continues to give the AI synthetic anchor the ability of body language, making the AI synthetic anchor more intelligent.
[13-14]
"AI synthetic anchor" has not only achieved technological innovation and breakthroughs in the field of global AI synthesis, but also pioneered the synthesis of real-time audio and video and AI real image in the field of news. "AI synthetic anchor" combined with Xinhua News Agency's "live news" report, the "AI synthetic anchor" and short video integration production.
[15]
The newly launched standing AI synthetic anchor is about to be put into the preparation of Xinhua News Agency's national two Sessions coverage in 2019, and will be in
Sogou search
, input method and other platforms to meet the user.
[16]
With the continuous improvement of the technical capability of "Sogou splitbody", the customization cycle of AI synthetic anchors is also reduced, and the synthesis effect and stability are also improved. With only a small amount of users' real audio and video data, a high-fidelity dope-body model can be customized to help media upgrade in the fields of media transformation, news timeliness, and cross-language communication ability.
[13]
↑ Report pages of BBC and REUTERS
Reference News reported
News anchors in China may face some new competition, according to foreign media reports
Artificial intelligence
Robots can imitate human facial expressions and mannerisms when broadcasting news.
[15]
In addition, The Times, Fox News, Russia Today TV, France 24 TV, Newsweek, the Los Angeles Times, National Public Radio and other media outlets also carried reports.
Reuters reported
Reuters reports that the AI anchor is modeled after Chinese news anchor Qiu Hao and wears a black suit and tie. This is part of China's efforts to boost its capabilities in artificial intelligence technology.
[15]
The Independent reported
The British "Independent" website said that the AI anchor of Xinhua News Agency is a
lifelike
The digital announcer can read text content by imitating the image and voice of a human anchor.
While being praised, some media believe that the "AI synthetic anchor" needs to be further improved. According to the US "Huffington Post" website, Xinhua seems to know that AI anchors need to be constantly improved, as the AI anchor emphasized in its first broadcast. It said: "As an AI news anchor in development, I know I have a lot to improve. Thank you."
[15]
Scholar's analysis
Britain
University of Sheffield
Artificial intelligence and
robotics
Professor Emeritus Noel Sharkey said that the AI anchor is a good experiment, and we will see it improve in the future.
