Research on the Innovation Trends of AIGC Technology Stacks Based on Patent Data
At the end of 2022, with the release of ChatGPT as a symbol of the emergence of the large language model, the wave of artificial intelligence (AI) science and technology innovation has rapidly pushed up people's perception of machine intelligence. Theoretically, we have experienced the shift from computational intelligence and perceptual intelligence, to the current cognitive intelligence, and the machine "autonomy" is constantly enhancing. In terms of technological innovation, with the empowerment of digital infrastructure elements, such as data, algorithms, and computational power, the release of the ChatGPT large language model and the Sora text-to-video model has become major milestones in AI innovation, pushing AI from the era of analytical AI into the new era of AI-generated content (AIGC).In the new era of AIGC, models have become repositories of knowledge and reshaped the intergenerational development of AI technology: before the advent of large models, general-purpose AI technologies such as deep learning, natural language processing, knowledge graph, intelligent speech, and computer vision have been accumulating and becoming increasingly mature; after the emergence of large models, large model technologies such as large language models and text-to-video models have been deeply integrated with general-purpose AI technologies, reaching new heights of universal and pervasive and creative and generative AI. This advancement has strengthened the connection between AI technology and data flywheel, driving the formation of a new paradigm of AIGC technology stacks characterized by computational power support, model leadership, and universal application.This paper focuses on the main research dimensions of China's AIGC patents, such as the change of patent quantity, the distribution of technology composition, and the analysis of value trend, and follows an analytical logic of the overall technology stack to the specific technology fields, and from the technology fields to the high-value technological innovations one by one. At the same time, recognizing the fact that AIGC technology is still developing at a high speed and its application is still in the early stage of exploration, this paper tries to design a model of innovation potential of key innovators from the three aspects of innovation activity, innovation scale, and efficacy distribution based on patent bibliographic data, in order to understand and forecast the future direction and development potential of AIGC technology. To analyze the application of AIGC technology stacks, this paper selects "smart industry" as a typical vertical scenario to analyze the patent innovation trend, technology composition, and high-value patents and study the development needs of AIGC technology stack application innovation. Finally, based on the interpretation and analysis of patent data, it summarizes the current innovation landscape and offers suggestions for the high-quality development of China's AIG technology stacks.
NETL
NSTL
万方数据
