In the field of artificial intelligence, a breakthrough has been made: the "sourceless water" problem of domestically produced large models has been solved.

In the field of artificial intelligence, a breakthrough has been made: the "sourceless water" problem of domestically produced large models has been solved.

At the "General Artificial Intelligence Industry Development Opportunities and Risks in the Big Model Era" forum at the 2023 World Artificial Intelligence Conference, a number of experts in the field of general artificial intelligence focused on large models and in-depth discussions on basic innovation, application technology and the future Prospects and other aspects of artificial intelligence issues.

Dai Qionghai, an academician of the Chinese Academy of Engineering, said: "Our country should deepen talent training and basic research on artificial intelligence in terms of policies, mechanisms and investment, strengthen original innovation, and avoid falling into the dilemma of 'water without a source'." He said in his keynote speech emphasized this point.

Wang Yu, tenured professor and chair of the Department of Electronic Engineering at Tsinghua University, pointed out that Shanghai already has many chip companies and algorithms, but how to deploy these algorithms on chips efficiently and uniformly is a very important issue. He emphasized that this is a key challenge faced by Shanghai in the field of artificial intelligence.

From the perspective of basic research, Dai Qionghai believes that my country’s breakthrough achievements in large-scale innovation are relatively limited. His point of view is that China’s talents in the field of artificial intelligence are mainly concentrated in applications, so there is huge development potential in application scenarios and technical levels. However, in terms of talent at the basic level, China is clearly at a disadvantage and lacks original innovation.

Dai Qionghai said that the innovative development of artificial intelligence requires three pillars: algorithms, data and computing power. Algorithms determine the level of intelligence, data determines the scope of intelligence, and computing power determines intelligence efficiency. Generally speaking, it is expected that in the next five years or so, large algorithm models will become the core basic platform for artificial intelligence applications.

Dai Qionghai also pointed out that brain intelligence is the new direction of the future. New artificial intelligence algorithms that integrate brain and cognition will lead the development of a new generation of intelligence. He suggested that the government encourage enterprises to lead the construction of large models, explore the combination of biological mechanisms and machine features, and further promote basic research and application development. He predicted that artificial intelligence with cognitive intelligence as its core will begin to be widely used in ten years.

In addition, Dai Qionghai also reminded people to be wary of security issues in large model applications. Large models are not yet capable of verifying the credibility of outputs, such as generating deceptive content. He emphasized that problems with large model applications are not as simple as computer network viruses. Once problems occur, they will have a disruptive impact. Therefore, safety and trustworthiness should be explicitly discussed during the application of large models.

It is necessary to focus on solving the pain points required to solve the four problems faced by the implementation of large-scale domestic models. First, the problem of long text processing needs to be solved. Secondly, the cost performance of large models needs to be improved. Third, large models need to be applied to multiple vertical domains. Finally, there is a new requirement for one-stop deployment. He emphasized that solving these needs will promote the development of the entire industry chain.

In the forum, participants put forward more opinions and suggestions on the development of large models. Some experts believe that dependence in the chip field can be compensated by enhancing the development and application of domestic large-power computing chips. They emphasized that although some chip companies have emerged in China, the ability to efficiently and uniformly deploy algorithms on chips needs to be further strengthened.

At the same time, experts also mentioned the application issues of large models in different vertical fields. In fields such as medical and finance, obtaining large-scale corpus data is a huge problem. Therefore, establishing a large model of a universal base and conducting detailed fine-tuning will help improve the basic performance of various industries.

It is generally believed that automating the deployment and optimization of large models into integrated solutions is an important trend. Improve overall efficiency and achieve more cost-effective results by implementing a layered approach to optimizing software and hardware collaboration, compilation optimization, and hardware infrastructure deployment. Experts call for further exploration of efficient fine-tuning algorithms to meet the needs of large models in different vertical fields.

Participants reached a consensus, emphasizing that the development of large models requires the joint efforts of governments, enterprises and academia. The government should strengthen policy guidance and promote basic research and talent training. Enterprises should assume a leading role and increase investment and promotion in the construction of large models. The academic community should strengthen cooperation with industry to promote the transformation and application of scientific and technological achievements.

Experts emphasized the need to strengthen research and exploration on security and credibility in the development of large-scale models. They advocate the establishment of corresponding norms and standards to ensure that the application of large models does not bring adverse effects and risks.

Finally, the participants expressed that the development of large models will bring huge opportunities to the artificial intelligence industry, but they also need to be alert to potential risks and challenges. They encourage all parties to conduct in-depth cooperation in the research and development, deployment and application of large models to jointly promote the healthy development of artificial intelligence and social progress.

Experts conducted extensive discussions and exchanges on the development and application of large-scale models at the forum of the World Artificial Intelligence Conference. They provided valuable insights and suggestions on basic innovation, technology applications and future prospects in the field of artificial intelligence, pointing out the direction for the development of the artificial intelligence industry.

The above is the detailed content of In the field of artificial intelligence, a breakthrough has been made: the "sourceless water" problem of domestically produced large models has been solved.. For more information, please follow other related articles on the PHP Chinese website!