On April 17, it was reported that the China AIGC Industry Summit hosted by qubits was held in Beijing today, gathering representatives of industry, academia, research and investment in the field to share and discuss hot topics and trends such as the new application paradigm led by generative AI and the digital transformation of industries. At the same time, the China AIGC Industry Panorama Report and the list of "AIGC Enterprises & Products to Watch in 2024" were also officially released at the meeting.
If in 2023, the large model reconstructs a new generative world with Prompt, then what we want to explore in this new world today is the new application based on the large model.
Based on such industry insights, with the theme of "Hello, New Applications", this year's China AIGC Industry Summit will comprehensively discuss and share the current situation, opportunities and future of China's AIGC industry from multiple dimensions such as technology, industrial landing, investment, and products, as well as industry chain players such as application layer, model layer, and infrastructure layer, as well as keynote speeches and roundtable discussions on the most cutting-edge technologies and industry insights.
Representatives of application layer players, including Meitu, Kingsoft, Evernote, Milestone Power, and Get, shared their latest AIGC application practices at the conference.
Including Microsoft, Kunlun Wanwei, Qualcomm, Xiaoice, Ant and other mainstream model layer players in the industry. The large models they launch are either generic or vertical, underpinning each AIGC application.
The infrastructure layer is also indispensable. Yang Fan, President of SenseTime's Large Device Business Group, Wang Xiaoye, Technical Director of Amazon Web Services China Product Department, and other corporate executives came to the scene to share their industry landing progress.
In addition to the pioneers of the AIGC industry wave, experts from the academic and investment circles brought industry insights from a more macro perspective, including Professor Lu Zhiwu of Hillhouse School of Artificial Intelligence of Chinese People's University, Yuan Li, Assistant Professor of Shenzhen Graduate School of Peking University, and Zhang Lu, a well-known investor in Silicon Valley.
A roundtable panel was set up in the China AIGC Industry Summit, which also gathered many domestic AIGC experts to have a heated discussion on the implementation and future trends of AIGC applications, including Xu Dong, head of Alibaba Cloud's Tongyi large model business, Gao Yushi, vice president of technology of Easy Group, and Zhou Jian, founder and CEO of Lanma Technology.
In addition, Meng Hong, founder and CEO of qubits, officially released the "China AIGC Industry Panorama Report" and the list of "AIGC Enterprises & Products to Watch in 2024" at the meeting.
Microsoft, Alibaba, Qualcomm and other players explore new AI applications
The first to open the speech was Tao Ran, Chief Operating Officer of Microsoft Greater China. When we mention the leader in the development of global AIGC applications, Microsoft is indispensable. In the past, we have witnessed the industry storm caused by Microsoft in many areas. The world's most advanced large models are also supported by Microsoft, and their capabilities have been continuously iterated and expanded.
Li Mian, General Manager of Microsoft Azure Business Unit in Greater China, shared how Microsoft Copilot and Azure AI Platform accelerate the implementation of enterprise applications in the AI era.
Focusing on the country, it is a prosperous scene of a hundred flowers.
For example, Kunlun Wanwei, the dark horse of the AIGC era, has carried out a lot of explorations in the past year, and recently there is a Tiangong music model that has attracted much attention in the industry, it is the only publicly available music AIGC product in China, and its AI vocal synthesis can reach the SOTA level in the industry.
Fang Han, chairman and CEO of Kunlun Wanwei, shared the process behind the creation of the Tiangong music model at the scene. With 29 years of experience in the Internet industry, he is one of the earliest promoters of Chinese Linux open source and one of the earliest network security experts in China.
In the past year, the Qianwen series of models has continued to be open-sourced, causing a lot of sensation in the global developer circle. The open source and openness of large models will also promote the emergence of new applications of AIGC. Lin Junyang, head of open source of Tongyi Qianwen, shared his views on the opening and closing sources of large models at the scene.
Lin Junyang is a senior algorithm expert at Alibaba, who has not only participated in the research and development, open source, and integration with external systems of the Tongyi Qianwen large model, but also participated in the creation of large models such as the ultra-large-scale pre-trained model series M6 and the general unified multi-modal pre-training model OFA.
In the "mold era" set off by the storm of large-scale model technology, the end products we use every day are also accelerating the change. In the new era, how can the terminal be AI? Qualcomm's choice is to start innovating from the underlying chip to promote the transformation of upper-layer application products.
Wan Weixing, head of Qualcomm AI product technology in China, revealed the new capabilities brought by large models on the terminal side. He is currently responsible for the planning of the software and hardware of the AI engine on the Qualcomm intelligent device side and the construction of related ecosystems. He has also participated in a number of research projects in the field of multimedia on Qualcomm mobile platforms.
Virtual digital humans are also a track that has been further developed under AIGC technology. In this field, Xiaoice is a company that started early and continues to follow up the development and iteration.
Xu Yuanchun, co-founder and chief operating officer of Xiaoice and head of the AI Creativity Lab, led the team in technology research and development, product innovation and application implementation in the field of AI creation. At the meeting, he shared the latest exploration of large models + virtual humans.
Among the productivity tools empowered by large models, programming tools are undoubtedly the one that has attracted the most attention. How far have AI programmers evolved now?
Li Jianguo, head of CodeFuse, Ant Group's code model, discussed the changes brought about by AI code generation. Li Jianguo holds a Ph.D. from Tsinghua University, and he has been deeply involved in the field of machine learning and deep learning for many years, and his papers have been cited more than 10,000 times. He previously worked at Intel, where he contributed important hardware features to CPUs from the 10th Gen Core onwards.
Meitu, WPS, Evernote, and others discussed the commercialization of AIGC applications
In addition to the active exploration of players in the model layer, there are also pioneer representatives who are quickly entering the commercialization of AIGC applications.
When it comes to AIGC, the hottest thing this year is multimodality, especially Wensheng videos like Sora, which have high hopes from insiders.
Last year, Meitu's visual model MiracleVision was unveiled, and Meitu also launched the "Image Productivity Family Bucket" represented by Meitu Design Studio, WHEE, Kaipai and other products, which has since opened the exploration and experimentation of China's imaging industry in the field of AI.
Wu Xinhong, founder, chairman and CEO of Meitu, used the Meitu model as a starting point to dig deep into the development path and application of the large model technology behind it, and the industry trend.
In the field of general office, Kingsoft Office announced the upgrade of WPS Office a few days ago, and also released WPS AI Enterprise Edition and WPS collaboration, opening up the three major capabilities of document, AI and collaboration. In this summit, Yao Dong, vice president of Kingsoft Office and general manager of the R&D middle office division, further gave an in-depth interpretation of this.
Yao Dong has been engaged in software development for nearly 20 years, and has presided over the development of multiple platform client-related technologies, and is currently responsible for the R&D middle platform, including R&D management, infrastructure, AI algorithm research, engineering implementation and related product and service development.
In addition to WPS, Evernote is also one of the most popular office products. Evernote's Chairman and CEO, Yi Tang, will share how ordinary people can use the power of AIGC to improve productivity.
Tang Yi has accumulated profound experience in the fields of technology entrepreneurship, multinational enterprise management, investment and financing. Now, under his leadership, Evernote has launched its self-developed lightweight large language model "Elephant GPT", and launched Impression AI products and a series of product functions, fully empowering its own software and hardware ecosystem.
In addition, in the field of embodied intelligence, thanks to the blessing of AIGC's capabilities, it has received unprecedented attention this year. Zhang Li, co-founder and chief operating officer of Zhuji Power, introduced the current status and industry implementation of general robots at the summit.
Zhang Li has been deeply involved in many industries such as autonomous driving, communications, and IT, and has served as the chief operating officer of China's leading autonomous driving companies, and also served as the senior vice president of Cisco Greater China.
Of course, in the face of ever-changing AI technology, learning has become a very important thing. How should ordinary people adapt and develop in the AI era, from learning AI to flexibly applying AI?
The co-founder and manager of the AI Learning Circle, Kuaidao Tsing Yi, shared at the meeting how to effectively improve personal abilities through AIGC. As a senior product manager, he led the "get" team to try in the field of AI, and developed AI applications such as "get learning assistant" and "start practicing".
SenseTime Amazon talks about AIGC's new infrastructure
Infrastructure is also indispensable in the era of large models, and computing power is the key.
Yang Fan, co-founder of SenseTime and president of the large device business group, will use SenseTime large device as an example to interpret the new way of computing power that continues to promote the application empowerment of artificial intelligence and the development of the ecosystem.
Yang Fan is responsible for the implementation of SenseTime's strategic business equipment, leading the construction of the "SenseTime Shanghai New Generation Artificial Intelligence Computing and Empowerment Platform", creating artificial intelligence infrastructure services, and continuously promoting the application empowerment and ecosystem development of artificial intelligence.
When it comes to computing power in the AIGC era, cloud computing is definitely an indispensable one.
Wang Xiaoye, Technical Director of the Product Department of Amazon Web Services China, a global cloud computing giant, shared generative AI practices and use cases.
Wang Xiaoye is responsible for leading a team of technical experts, providing technical practice guidance for cloud products and open source solutions, and is committed to the application and promotion of Amazon Web Services full-stack services in the Chinese market.
AIGC Insight
The development of the industry also needs to be viewed from a macro perspective, including the market, including the academic community.
In Silicon Valley, where has the generative AI industry ecosystem evolved? What new opportunities are Silicon Valley entrepreneurs paying attention to? Zhang Lu, a well-known investor in Silicon Valley and founder of the technology VC Fusion Fund, shared her judgment on the latest industry trends.
A graduate of Stanford University and now managing nearly $400 million in capital, focusing on investments in healthcare, AI and deep technology, she once asserted that the scale of the industrial innovation opportunity is 10 times larger than that in the Internet era.
Before ChatGPT, the team of Lu Zhiwu, professor and doctoral supervisor of Hillhouse School of Artificial Intelligence at Renmin University of Chinese, designed Wenlan, a general multimodal pre-trained model for Chinese, and published it in Nature Communications. Then, last year, it was the first in the industry to launch ChatLMG, a multi-modal dialogue model, and VDT, a Sora-like video generation base model.
In this summit, he made an in-depth analysis of VDT and decrypted the most popular Wensheng video technology.
In the application of AIGC, the work of the Peking University team has repeatedly appeared in hot searches, and vertical products such as ChatExcel and ChatLaw have attracted much attention. Yuan Li, the team leader behind the scenes, assistant professor of Peking University Shenzhen Graduate School, came to the scene to share the exploration process behind it.
Yuan Li focuses on the research direction of multimodal deep learning, and a single paper has been cited more than 1,000 times, and the team recently launched the Sora reproduction project.
Roundtable Panel: AIGC Application Landing and Future Trends
At the end of the summit, a roundtable panel was also set up, which also gathered several domestic AIGC experts to discuss the implementation and future trends of AIGC applications.
Xu Dong, head of Alibaba Cloud's Tongyi large model business, Gao Yushi, vice president of technology of Easy Group, and Zhou Jian, founder and CEO of Lanma Technology.
Xu Dong has served as the general manager of Alibaba Cloud South China, the general manager of the Internet industry, and the head of Wuying products, and has rich product practice and business experience in the fields of cloud native, device-cloud architecture, and AI large models.
Gao Yushi led the R&D and construction of "AICare", an AI intelligent system in the field of health security, which is based on large model technology and widely used in various business scenarios of Easy Group, comprehensively improving the intelligence level of various business scenarios of the Group.
Zhou Jian graduated from Shanghai Jiaotong University with a bachelor's degree and master's degree, won the ACM World Championship in 2002, and has ten years of continuous entrepreneurial experience.
The list of China's most noteworthy enterprises & products of AIGC has been announced!
2024 is the "first year of AIGC application landing", what new applications, new forms, and new species are worth looking forward to, and what qualities can enterprises enjoy the dividends of the times?
A reference answer has been released, and the results of "AIGC companies & products worth paying attention to in 2024" have been announced, with 50 AIGC companies and 40 AIGC products selected. For the full list, please refer to the qubit WeChat public account.
In addition, Meng Hong, the founder and CEO of qubits, also released the "China AIGC Application Panorama Report" on the spot, and the first AIGC application panorama was released.
The report also covers the current state of the market, the distribution of players/products, the future development stage, and the market size.
The qubit think tank predicts that the mainland AIGC application market will reach 20 billion in 2024. In 2030, the application of AIGC in mainland China will become a trillion-dollar market, with an average compound growth rate of more than 30% in five years (2024-2028).