Recently, large language models (LLM) have attracted attention with the release of ChatGPT, a user-centered chatbot released by OpenAI. Whether a large model is good or bad mainly involves three aspects: algorithm, computing power and data. The algorithm determines its own learning ability, computing power affects the learning efficiency of the algorithm, and data determines the training effect of AI.
Artificial intelligence (AI) is one of the core driving forces of industrial development and a key factor in promoting the integration of emerging technologies such as graphics processing units, the Internet of Things, cloud computing, and blockchain with the new generation of big data and Industry 4.0. A multi-angle systematic analysis of artificial intelligence from underlying mechanisms to practical applications, from basic algorithms to industrial results, and from current status to future trends. Although there are still many problems with artificial intelligence, there is no doubt that artificial intelligence has become an innovative and revolutionary assistant in a wide range of applications and fields.
Artificial intelligence is an important driving force for a new round of technological revolution and industrial transformation. In recent years, my country's artificial intelligence industry has made positive progress in technological innovation, industrial ecology, integrated applications, etc., and has entered the first echelon in the world. Big data and big computing power have pushed the development of artificial intelligence toward the era of big models. AI large models refer to models that have been trained on large-scale data and can be adapted to a wide range of downstream tasks after fine-tuning. As the scale of parameters continues to expand, large AI models continue to emerge with new capabilities in fields such as language, vision, reasoning, and human-computer interaction.
At present, the number of large models developed by China and the United States accounts for more than 80% of the global total, and China ranks second only to the United States in terms of the number of large models. my country's early deployments in the field of artificial intelligence have laid a solid foundation for the development of large models. It has established systematic research and development capabilities covering theoretical methods and software and hardware technologies, forming a large model technology group that keeps pace with the world's cutting-edge. Data shows that as of 2023, the total number of large models in my country with a parameter scale of more than 1 billion will reach 79.
The United States leads the way, accounting for 53% of investment in AI companies ($27 billion), followed by China, accounting for 10%
Industry development prospects:
1. Policies support industry development
Artificial intelligence is an important part of national strategy, the focus of future international competition and the new engine of economic development. In recent years, China's artificial intelligence industry has received great attention from governments at all levels and key support from national industrial policies. The country has successively issued a number of policies to encourage the development and innovation of the artificial intelligence industry.
In April this year, the Cyberspace Administration of China released the "Measures for the Administration of Generative Artificial Intelligence Services (Draft for Comments)", which for the first time clarified the legal responsibilities and duties of generative artificial intelligence "providers" in terms of content production, data protection, privacy security, etc. The legal basis establishes the safety assessment regulations and management methods for artificial intelligence products. Recently, the Cyberspace Administration of China, together with the National Development and Reform Commission, the Ministry of Education, the Ministry of Science and Technology, the Ministry of Industry and Information Technology, the Ministry of Public Security, and the State Administration of Radio, Film and Television, announced the "Interim Measures for the Management of Generative Artificial Intelligence Services", which will be effective from August 15, 2023. .
2. Growth in scale of AI basic data services
In the AI industry chain, algorithms, computing power and data together constitute the three core elements of technological development. The implementation process of algorithm models from technical theory to application practice relies on a large amount of training data. The more, more complete, and higher quality the training data is, the more reliable the conclusions inferred by the model will be. With the development of emerging technologies such as artificial intelligence, blockchain, and IoT, the scale of global basic data services is growing rapidly. In the future, the realization of complex intelligent scenarios such as intelligent manufacturing, metaverse, and generative AI will put forward higher requirements for basic artificial intelligence data services.
3. Accelerate the construction of computing power infrastructure
To develop large AI models, computing power is an important support. It is crucial to speed up the construction of computing power infrastructure, optimize the layout of computing power resources, and improve the intensity of computing power application. With the deployment of national hub nodes of the national integrated computing power network and the advancement of the "Eastern Digital and Western Calculation" project, the construction of my country's computing power infrastructure has developed rapidly and the scale of computing power has continued to expand. According to the "Guidelines for the Innovation and Development of Intelligent Computing Centers" jointly issued by the National Information Center and Inspur Information, more than 30 cities across the country are currently building or proposing to build intelligent computing centers. During the "14th Five-Year Plan" period, investment in intelligent computing centers can drive The core industry of artificial intelligence has grown by approximately 2.9-3.4 times.
Summarize:
AI technology promotes the development of artificial intelligence-assisted industries in all walks of life. However, challenges remain. Work is currently underway to address multimodal data integration, balancing quantitative algorithm performance and qualitative model interpretability, and protecting model security, learning, and model bias.