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

"generative artificial intelligence is the most important technological revolution in the past 40 years," is the latest research by Microsoft co-founder Bill Gates. When intelligence emerges, how to promote scientific research and applied innovation has become the focus of the industry.

On September 7, at the sub-forum of the 2023 Bund Conference "New Generation data Base-exploring the Application and Development of Graph Intelligence", Ant Group brought a fusion research-"Large Graph Model" (LGM). This study combines graph computing with graph learning and large language models, and makes use of the generation ability of large language models and the relevance analysis ability of graph computing to provide more intuitive and comprehensive information presentation and more accurate insight. in order to better solve the massive and complex digital application problems. At present, Ant has completed the first phase of "generative heterograph enhancement", and the related papers have been included by the World computer Association (WWW 2023).

Graph computing is a powerful data processing technology, which can solve the relevance problems in complex relationship networks. it is used in the fields of financial anti-fraud, weather forecast, drug development and even brain-like research. Large model is the most likely technology to move towards general artificial intelligence, and its tasks are comparable to or even better than those of human beings in some fields.

Why use cutting-edge technology to drive cutting-edge technology? Can't large language models accomplish data analysis and mining tasks independently? Liu Yongchao, a senior technical expert at Ant Group, said that the large language model can infer implicit relationships, but it is impossible to draw diagrams, while the study of data relationships needs to reflect clear links, and it is easier to understand by using graph structures. " To combine large language models with graph computing is to first reason logically from huge amounts of information, and then use supercomputing to work out relationships, which is similar to connecting a supercomputer to the human brain, which makes it more powerful, "explains Liu Yung-Chao.

▲ Ant Group Liu Yongchao shares the research results of "Big Picture Model" in this study, Ant Group has mainly carried out two tasks. First, the large language model is used to enrich graph data. Different from the ordinary context-dependent model, the large language model can generate new data points based on existing data. This work is called "generative heterogeneous Graph Enhancement", which extends and enriches different types of graph data through large language models. Second, use Prompt (an instruction or prompt) to guide the model to learn and discover specific data characteristics. For example, if you set a Prompt "common characteristics of groups that default more than three times a year", the model will generate data samples that meet specific conditions. This capability can speed up the process of data analysis and feature discovery.

Ant Group is a leader in the field of graph computing. TuGraph, a graph computing platform jointly developed with Tsinghua University, has broken the LDBC SNB world record for authoritative evaluation of graph database for three times, won the "leading Scientific and technological Achievement" Award of the World Internet Conference in 2021, and was selected as the "leader" quadrant of IDC MarketScape China Map database market in 2023. In recent years, the industry has carried out a variety of attempts to map intelligence to general artificial intelligence. "Big picture model" is the research that ants began to study in order to solve industry problems at the beginning of 2022. With the outbreak of the large model at the end of 2022, the feasibility of this study has been verified.

At this forum, experts and scholars in the field of map intelligence at home and abroad also shared more hot developments. M. Tamer ö zsu, a professor at the University of Waterloo in Canada, shared the challenging flow graph computing techniques in the industry. Chen Huajun, a professor at the School of computer Science of Zhejiang University, talked about the opportunities and challenges of knowledge processing in the era of large models. Chen Hongyang, deputy director of the Map Computing Research Center of Zhijiang Laboratory, brought the latest research on scientific computing and biomedical research and development of Zhijiang Vermilion Bird diagram. Li Ya, co-founder and deputy editor of Machine Heart, believes that the research on the combination of graph intelligence and large models is expected to bring a significant improvement to data intelligence.

With regard to the possibility of artificial intelligence and graph computing, the Bund Map Intelligence Forum brings together views from different perspectives and depicts an important development context for the development of map intelligence.

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.