Shulou(Shulou.com)11/24 Report--

Artificial intelligence, generate the future! On August 25, 2023, the first International Conference on artificial Intelligence generated content (The 1st International Conference on AI-generated Content, AIGC 2023) successfully opened in Shanghai. This is the first international conference focusing on generative artificial intelligence in the world. At the moment when large models detonate the upsurge of generative artificial intelligence, the holding of this Congress has contributed a gluttonous feast for academia and industry.

▲ 's first International Conference on content Generation of artificial Intelligence, is General artificial Intelligence "Universal"? What are the problems that need to be solved urgently in the big model and the current deep learning? What are the other scientific challenges of large language models? How to solve the problems brought by the large model, such as invasion of privacy, discrimination, security and deep forgery? All this can be heard at the AIGC 2023 conference. From August 25 to 26, AIGC 2023 was held at Radisson Hotel on the Yangtze River, Shanghai, China. Many academicians, experts, scholars and professors at home and abroad, as well as enterprise representatives and professionals in the field of artificial intelligence gathered together to share cutting-edge academic achievements, collide ideas and focus on industrial landing applications. This conference is hosted by the Organizing Committee of AIGC 2023. Professor Qiu Xipeng, head of the computer School of Fudan University and MOSS system of Fudan University, served as the chairman of the conference; Professor Ouyang Wanli, leading scientist of Shanghai artificial Intelligence Laboratory, and Associate Professor Li Hongsheng of the Chinese University of Hong Kong co-chaired the conference; Professor Zhao Feng, director of the University of Science and Technology of China and National Engineering Laboratory of brain-like Intelligence and Application, served as chairman of the Program Committee of the Conference. Professor Miao Zhengqian of Tongji University and Professor Yan Junchi of the computer Department of Shanghai Jiaotong University co-chaired the Technical Committee. The AIGC2023 conference is hosted by Sibaicheng Technology and Spectrum Academy, and co-organized by Yiou. At the same time, it is supported by experts from the Institute of Science and Chemistry of Japan, the University of Hong Kong, Durham University of the United Kingdom, Nanyang University of Singapore, Shandong University, Jinan University, Fudan University, Tongji University, Shanghai Jiaotong University, East China normal University, Shanghai artificial Intelligence Laboratory, Chinese University of Hong Kong, Hong Kong Baptist University and University of Toronto. On the two-day agenda The conference has set up a plenary meeting (main forum), a sub-forum on industrial practice in the AIGC era, a sub-forum on innovation and application of AIGC, a special forum on digital transformation and landing of large models, a symposium on basic models and fine-tuning strategies (held in Japan), a symposium on trusted basic models under imperfect data (held in Hong Kong), a symposium on reinforcement learning in the real world (held in the United States), and keynote reports. A variety of academic activities and links, such as invited reports, oral reports, wall reports and so on.

Professor Zhao Feng, Chairman of the Program Committee of the ▲ Conference, presided over the opening ceremony and plenary meeting of the Conference. Professor Zhao Feng, Chairman of the Program Committee, presided over the opening ceremony. Professor Zhao Feng pointed out that the AIGC application ChatGPT, which was launched in November 2022, quickly gained a large number of users by virtue of its excellent performance in the fields of semantic understanding, text creation, coding, logical reasoning and knowledge question and answer, and the research and application related to AIGC also developed at an amazing speed. The iterative speed shows an exponential burst. Holding AIGC international conference under this background is to promote the exchange and cooperation of artificial intelligence technology and large model research in China and expand international influence, and effectively promote the deep integration of AIGC and industry. He hopes that the experts can share their insights and visions in the field with you at AIGC 2023, so as to show you a wonderful knowledge sharing.

Professor Qiu Xipeng, Chairman of ▲ Conference, delivered opening speech to AIGC 2023. Professor Qiu said in his opening speech that today, with the rapid development of data mining, algorithms, computing power and other related technologies, AIGC has shown strong creativity and expression, and is about to release a new revolution in content production. Technological breakthroughs to achieve applied innovation, AIGC has been closely related to medical care, manufacturing, transportation, entertainment, finance, education and other fields. It not only changes our way of life, but also opens up commercial opportunities, but also provides new opportunities for the development of various industries. However, the development of AIGC is also faced with a series of problems and challenges: how to ensure the quality and credibility of its generated content? How to solve the problem of data scarcity and sample deviation? How to balance the relationship between creativity and copyright protection? These are all things we need to think about, explore and solve. Through the AIGC conference, we hope to encourage academia and enterprises around the world to strengthen cooperation, seize opportunities, meet challenges, and provide strong support for promoting the innovative application and practice of AIGC.

Academician Jiao Licheng of ▲ released the report of "General artificial Intelligence: some problems of large models and Deep Learning". Professor Jiao Licheng, academician of European Academy of Sciences, IEEE Fellow, Huashan Outstanding Professor of Xi'an University of Electronic Science and Technology and Dean of artificial Intelligence Research Institute, delivered a report entitled "General artificial Intelligence: some problems of large models and Deep Learning". Academician Jiao pointed out that the big model has brought us changes in world outlook, philosophy, thinking and, more importantly, the research methodology of generative artificial intelligence. Up to now, the big model has the beginning of industry, a good start, and everyone's enthusiasm to study it together, but its development is still on the way. There are still challenges in theory, method, scenario, as well as its own key core technologies. Academician Jiao summed up 10 challenges facing ChatGPT: not yet passing the Turing test, mathematical foundation, mathematical logic ability to be strengthened, limitations of code generation, semantic perception and cognition, model bias, speech recognition problems, weak environmental awareness, multitasking, emotional and emotional intelligence, and interpretability and risk. Finally, Academician Jiao concluded that the development of deep neural networks and large model algorithms should not only pursue more, fast and good, but only driven by data, model and knowledge can solve the above challenges in their development. The convenience brought to us by deep learning and generative artificial intelligence is unprecedented. similarly, the challenges faced by its security, interpretability and ethics are also unprecedented. Only by dealing with this contradiction can we make the field of generative artificial intelligence develop healthily, greenly and rapidly.

Professor Qiu Xipeng, ▲, published the conference report "Scientific challenges of large-scale language models". As the first person in charge of the conversational large-scale language model MOSS system in China, the conference report of Professor Qiu Xipeng, School of computer Science, Fudan University has attracted much attention. Professor Qiu pointed out in his report entitled "Scientific challenges of large language models" that although natural language processing has entered the era of large models, however, the large model brings many problems, such as serious homogenization, track congestion, returning to feature engineering, high computing power, high energy consumption, and difficulties in evaluation. Professor Qiu put forward ten scientific challenges for large-scale language models at the conference: what is the architectural design of further expanding the scale of the model, the source of the ability of the thinking chain, and what is the pre-training goal of the large model? Common defects of generating models-"hallucinations", multimodal extension, continuous learning and knowledge sources of large models, real-time learning, from dialogue to agents, automatic evaluation and popularization of large models. In his report, Professor Qiu introduces his own exploration path and experience in scientific research practice in view of these scientific challenges. Professor Qiu concluded that the large model is not only an engineering problem. In addition to the great demand for computing power, the general large model still has many challenges in terms of credibility and security, and major scientific research problems need to be tackled. Solve scientific problems such as intelligent emergence, complex reasoning, efficient architecture, knowledge fusion, learning strategies and other scientific problems, and provide theoretical and technical support for the landing of general large models.

Professor ▲ Maria Gini delivered a report on "call to artificial Intelligence researchers: cracking down on Deep counterfeiting". Professor Maria Gini, a distinguished professor at the University of Minnesota and the former chairman of IEEE Fellow, ACM Fellow, AAAI Fellow and ACM SIGAI, shared her concerns about the large model from another perspective in a speech entitled "call to artificial Intelligence researchers: fighting Deep counterfeiting" Professor Maria Gini expressed that with the rapid popularity of large-scale language models, it brings new opportunities as well as new challenges. Artificial intelligence researchers should find ways to prevent the problems caused by these models, such as invasion of privacy, discrimination, lack of security and deep forgery. Some problems require the introduction of artificial intelligence system regulations, but some problems, such as deep forgery, also need technical solutions, which requires the participation of artificial intelligence researchers. Professor Maria Gini said: "as AI researchers, we have created AI technology, and we are duty-bound to participate in solving the privacy, security and other issues brought about by AI."

Keynote report of the first ▲ International Conference on artificial Intelligence generated content Professor Xiao Qinglong, Chairman and Chair Professor of Information Systems, City University of Hong Kong, Professor Zhu Jianke, School of computer Science and Technology, Zhejiang University, Professor Li Xiuxian, leading talent of Tongji University and Shanghai, Professor Xiong Deyi, head of Intelligence and Computing Department and Natural language processing Laboratory of Tianjin University Many experts, such as Cui Wanyun, Associate Professor, School of Information, Shanghai University of Finance and Economics, attended the conference and delivered academic speeches.

AIGC, the first ▲ International Conference on content generated by artificial Intelligence, has made remarkable achievements in industrial applications in some fields, but there is still a broad space for industrial applications in the future. With regard to the industrial application of AIGC, Dr. Bo Liefeng, head of the XR Laboratory (X Lab) of Alibaba Dama Hospital, Mr. Miao Guocheng, director of Yiou Company and general manager of Shanghai Company, Dr. Jiang Yahong, founder and CEO of Youlian Times. Shushan Lu, founder and chief researcher of careful thinking Society, and Han Yi, founder and CEO of Titan Mathematics, respectively gave speeches on digital people and 3D content generation, the prospects and challenges of AIGC, the investment password for the value of AI scenes, the creation and application of human 3D digital avatars, the trend of AI applications and the future of AI-driven industry. In the simultaneous "AIGC Innovation and Application Forum", Chen Ye, founder and CEO of Tiger Technology, and Microsoft experts Twelve artificial intelligence founders, CEO and leaders, including Peng Aihua, developer of Huawei Cloud, Li Yuan, Senior Vice President of Zhiyun, and Xu Hao, General Manager of Solutions Department of Kingdee Software (China) Co., Ltd., held in-depth industry sharing and wonderful discussions on the research and development of large models, practical application of AIGC, AIGC and enterprise services, AIGC and future collaboration.

The first ▲ International Conference on artificial Intelligence generated content AIGC Innovation and Application Forum on August 26, the AIGC Conference held a special international academic conference. Yu Mingyu, Director of the Institute of artificial Intelligence and Marine Robotics of Dalian Maritime University, Dr. Zhang Jingfeng, Department of computer Science, University of Auckland, and Dr. Han Bo, Assistant Professor, Department of computer Science, Hong Kong Baptist University. Dr. Haipeng Chen, Assistant Professor, Department of data and Science, College of William and Mary, Dr. Yuxi Li, founder of RL4RealLife, Dr. Ruibo Liu, Google DeepMind research scientist, Dr. Hyunwoo Kim, Senior Research expert in Zhijiang Laboratory, Associate Professor Chen Jingjing, School of Information and Technology, Fudan University, Dr. Nayyar Zaidi, Network elasticity and Trust Center, Deakin University, Assistant Professor Dr. Zheng Yuan, University of Pittsburg Ryan (Zheyuan) Shi, Wu Jiaosu, member of the Standards Group of the artificial Intelligence Alliance of the Chinese Academy of Sciences, Luziwei Leng, Chief engineer of Huawei ACS Lab, and Xiaogang Wang, Senior researcher of LIGHTSPEED STUDIOS, gave lectures on the alignment and trend of large language models, decentralized online optimization and online games, neural reconstruction and rendering of AIGC, how to use large language models to optimize domain decision-making, perceiving the development and prospects of robots, and large-scale communication games. During the oral presentation session of the meeting, teachers and student writers from Xi'an University of Technology, Zhejiang Institute of Communication, Yunlin University of Science and Technology, Xi'an Jiaotong University, Macau University, Qingdao Binhai College, Zhujiang College of South China Agricultural University, SUNY Old Westbury and other universities shared their brilliant and rich academic achievements.

On August 26, the first ▲ International Conference on artificial Intelligence generated content, a special forum on "how to better meet the coming AGI era" was held at AIGC 2023. The forum is sponsored by the Chinese computer Society, hosted by CCF YOCSEF Shanghai academic Committee, co-sponsored by Intel (China) Co., Ltd., AIGC 2023 Organizing Committee, and supported by Unicom (Shanghai) Industry Internet Co., Ltd. This forum is divided into two parts: the morning public forum "Digital Transformation and the landing of large models" and the afternoon closed-door discussion session, and invited experts and scholars from many different fields to have an in-depth discussion on the upcoming AGI era from many angles, including how AGI technology will completely change our way of life, production relations and even social structure. What kind of preparations do various industries need to make? Does digital transformation provide enough training data for large models? How can we make good use of the opportunity of this technological revolution? From the point of view of large model data and industry digitization, this forum deeply discusses the impact of this technological innovation on society and the next specific measures, and discusses the direction of secondary innovation.

▲ morning open forum "Digital Transformation and landing of large models"

▲ afternoon closed meeting this International Conference on artificial Intelligence generated content (AIGC 2023) is not only an important event in the field of artificial intelligence generated content, but also a milestone in promoting the innovative development of AI. In this field full of opportunities and challenges, the vast number of industry experts and practitioners will continue to promote the innovation of artificial intelligence content generation technology, develop the unlimited potential of AIGC, and open up a new height for promoting the innovative application of AI technology. It is believed that in the future, AIGC will play a role in more fields and make greater contributions to social progress and human well-being.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.