Home>>Top

The hottest Jinshan cloud AI Development Forum was

2022-08-21

Detail

Jinshan cloud AI Development Forum was successfully held. AI broadened the boundaries of the video industry. On November 30, the "Ai empowering overlooking the horizon" AI application development forum hosted by Jinshan cloud was successfully held in Chengdu, Sichuan. As an important part of the 6th China network audio visual conference, this Forum gathered many top experts, scholars and talents in the industry. Once the news was released, it immediately attracted a large number of visitors to sign up. During the live broadcast of the forum, many friends thought and discussed, forwarded reports, liked, or left messages to express support. At the same time, the forum also attracted the high attention of dozens of well-known media

around the forum theme and "when video meets AI, what sparks will it spark?" This is a hot topic that the industry pays most attention to. Zong Jie, general manager of Jinshan cloud CDN and video cloud product center, an Zhuyong, deputy general manager of China digital media Internet technology department, Han Yi, deputy general manager of technology business group of CCTV International Network Co., Ltd., Hou Ximing, chief product architect of sound, Wang Yujun, director of human intelligence and cloud platform voice of Xiaomi, and Su Chi, chief algorithm architect of Jinshan cloud AI, six heavyweight guests expressed their views on the forum, Shared wonderful ideas. The atmosphere of the whole meeting was lively and enthusiastic, which was highly praised by the guests and audience

Zong Jie, general manager of Jinshan cloud CDN and video cloud product center, delivered a keynote speech

this forum has a clear theme and full content of "dry goods", and the guests' Multi-dimensional views and discussions are the biggest highlights. Zong Jie, general manager of Jinshan cloud CDN and video cloud product center, showed the audience a picture of the future of video cloud. Zong Jie said that with the video industry showing four trends: live interaction, artificial intelligence, ar/vr and HD, video cloud services will show four trends: deep marginalization of transmission, full cloud of video business, intelligence and edge computing. Jinshan cloud will follow the trend and create better products and services to empower more application fields through deep cultivation in the video industry

An Zhuyong, deputy general manager of the Internet Technology Department of Huashu media, delivered a keynote speech. An Zhuyong, deputy general manager of the Internet Technology Department of Huashu media, showed the application of AI in video content. An Zhuyong said that AI can do many things that were difficult to do before, and even require a lot of manpower. The biggest change it has brought to the video industry is the analysis of video content, which is a revolutionary breakthrough. For example, content auditing and smart tags can be achieved through artificial intelligence technology, "we stand on the shoulders of giants such as Kingsoft cloud and apply their AI technology to do our services well". He took the architecture of Huashu Ott broadcast control platform as an example, in which AI technology was used in the editing system, big data system, smart tag system, content review platform and smart recommendation, which greatly improved the work efficiency. In addition, it also realizes the intelligent voice control of Internet TV, and realizes user analysis and portrait through algorithm, which can better meet the needs of users

Han Yi, deputy general manager of the technology business group of CCTV International Network Co., Ltd., delivered a keynote speech

Han Yi, deputy general manager of the technology business group of CCTV International Network Co., Ltd., shared with the audience CCTV's practices and ideas of using artificial intelligence to achieve business upgrading. She cited the live broadcast of this year's World Cup in Russia as an example. CCTV reported that the live broadcast of the world cup had the highest number of simultaneous users of 6.58 million, which is currently the second highest concurrent record in the world. In the face of peak traffic, CCTV's intelligent scheduling system ensures that the node bandwidth is balanced and stable at the average water level through real-time detection and analysis of node bandwidth usage. In terms of HD video broadcasting, CCTV has greatly improved the visual experience of human vision with intelligent scene recognition, image diagnosis technology, AI processing and intelligent coding. Through intelligent processing and big data processing, CCTV will accumulate nearly 40 years of historical materials to revitalize resources, produce nearly 100 refined videos, and obtain more than 20million views

Houximing, the chief product architect of Sonora, delivered a keynote speech. Houximing, the chief product architect of Sonora, told you in his speech how artificial intelligence and RTC (real-time communication) technology will bring users a new experience. First, AI can improve the end-to-end audio and video quality of RTC. In social networking, education and other application scenarios, after video decoding, clearer effects can be obtained through AI based video hypersegmentation technology. Improve speech quality through algorithm and pattern recognition. The "crystal ball" product launched by sound also carries out quality analysis with the help of AI, which can carry out quality analysis and feedback on all real-time calls on the cloud, and help customers better deal with business logic on the operation side. Second, AI increases the real-time interactive experience of RTC. The most representative ones are face recognition and anthropomorphic voice change. In the application scenarios of social entertainment and education, users can use AI based beauty, mask changing, emotion recognition and other functions. Third, ai+rtc can achieve better business closed-loop. With AI, the video platform can not only realize the supervision of video and voice through video and voice porn detection, but also realize business value-added

Xiaomi AI and cloud platform voice director Wang Yujun delivered a keynote speech

Xiaomi AI and cloud platform voice director Wang Yujun shared "voice technology in Xiaomi" with the audience in his speech. Wang Yujun said that the research and application of intelligent voice technology is very important for Xiaomi. Share two figures: 8billion and 34million. 8billion is the cumulative number of times it has been awakened with the words "Xiaoai classmate" through Xiaomi smart speaker since March last year; 34million are the monthly active users of AI intelligent voice assistant Xiaoai, which provides Xiaomi with a large number of user scenarios and data. Xiaomi set up a speech group in February last year. The first is to do speech recognition. It made its first voice in the speech industry last June. In the next few months, Xiaomi combined with the data characteristics, constantly iterated the online speech recognition data model, so that this business grew rapidly. At present, it has realized speech recognition, speech wake-up, speaker recognition, speech synthesis and other functions on a variety of intelligent devices, as well as on the difficulties of end-to-end wake-up, noise suppression, speaker adaptation and so on, and made a technological breakthrough with the current frequency quadrupling technology

Su Chi, chief algorithm architect of Kingsoft cloud AI, delivered a keynote speech.

Su Chi, chief algorithm architect of Kingsoft cloud AI, brought the keynote speech "application practice of artificial intelligence in content security scenarios". Su Chi said that content audit faced problems such as unbalanced data and inconsistent labeling rules. The "golden eye" scientific team has designed a set of fine-grained cascade classifier (fine-graine requires the cooperation of manufacturers to solve d cascade model) solution, which solves these two problems. To solve the problem of data imbalance, cascade model is used for layer by layer filtering, which can significantly improve the response speed and concurrency, while taking into account the accuracy. Through the two steps of prediction of fine-grained labels at the bottom and dynamic adjustment of label mapping rules at the top, the problem of inconsistent label rules is effectively solved. Su Chi said that as an AI content service provider, Kingsoft cloud effectively connects its scientific research direction with the needs of regulators and content platforms, and provides customized solutions for different types and styles of video content

in the round table discussion session of the forum, the host had an in-depth discussion with five guests

in the round table discussion session of the forum, Tang cunchen, the host of the forum and professor of Wuhan University, had an in-depth discussion with five guests, including Zong Jie, an Zhuyong, Hou Ximing, Wang Yujun and Su Chi, on topics related to the video and artificial intelligence industry. The guests agreed that traditional industries are being pushed down and rebuilt, the general pattern of the Internet industry has basically taken shape, new industries are in the ascendant, and artificial intelligence will enter all aspects of people's life and work. Although the AI industry is still in its infancy, under the impact of the new wave of scientific and technological revolution, AI will soon be everywhere and omnipotent

from the wonderful speeches of all guests at this forum, we can deeply feel the current situation and development prospects of video +ai. Under the strong traction of artificial intelligence, the video industry has entered the era of intelligence in December 2013, and has shown an explosive trend. With the popularization of new technologies, the video industry is no longer satisfied with the existing intelligent technology, but seeks newer algorithms and richer applications, and strives to make more breakthroughs in the breadth and depth of applications. In the future, video will blossom in more industries and usher in a broader space for development

in the warm applause, the "Ai enabled but no matter the skill level and skill innovation can overlook the horizon" artificial intelligence application development forum came to a successful conclusion. The audience at the meeting said that the wonderful sharing of the guests had brought more new thoughts and provided many useful views and suggestions for the healthy and rapid development of the video industry. They hoped that such a meeting would continue to be held and bring more value