From July 6th to 8th, the 2023 World Artificial Intelligence Conference WAIC was held in Shanghai. As a top feast in the field of artificial intelligence, WAIC focuses on the development of general artificial intelligence, and brings together large-scale model star teams at home and abroad to jointly explore new formats of future industries. Data intelligence service provider Dipu Technology participated in the "AI Business Landing Forum" of the conference, and introduced the recent exploration results and industrial practice cases in the field of large-scale industry models.
Focusing on the landing value of large-scale industry models, the FastAGI platform empowers industrial innovation
Hu Xin, COO of Dipu Technology, attended the conference and said in his keynote speech: Focusing on the B-end business, Dipu Technology is based on the leadership of data intelligence technology, rapidly invests in large-scale industry model infrastructure and technological innovation, and implements the large-scale model To the key areas of enterprise intelligent decision-making and operation.
In recent years, Dipu Technology has been adhering to the core strategy of "Data+AI", and relying on the core product intelligent warehouse platform FastData, it has served the construction of data asset platforms for a large number of leading enterprises. With the rapid development of large-scale model technology, Dipu Technology has built the FastAGI platform this year. Through the practice of large-scale models in the field of enterprise services, it has precipitated the real business development needs of enterprises, and further created a large-scale model of the Dipu Technology industry, realizing from data-driven to intelligent Driver upgrades.
Accelerate industry landing: make industry big models become business intelligent decision-making and operation assistants
01 Based on the large model, upgrade the intelligence of Belle fashion data business scenarios
Belle Fashion is an in-depth partner of Dipu Technology in the construction of data platforms. The two parties have jointly built a number of basic data platform projects such as data lakes, brand brains, data management, and data intelligent applications, building a solid foundation for the digital transformation of Belle Fashion. data base. On this basis, the two parties continued to explore the implementation of large-scale models in the field of enterprise services this year, and have achieved preliminary results:
1.1 Store operation scenario model "Barley" - business analysis intelligent assistant
The data warehouse robot "Balley" can realize multiple rounds of interactive analysis of store business through the chatbox. Based on the industry large model trained by FastAGI through store operation analysis knowledge and the data lake built by Belle based on FastData, it can support store operation analysis and decision-making in all complex scenarios. In addition, the data warehouse robot also has text-to-SQL/API capabilities, supports intelligent conversational experience and reasoning, and enables business personnel to analyze store sales through various business scenarios.
1.2 Commodity Operation Scenario Model "Liying"—DeInsight
Another attempt by both parties in the field of large-scale models is to upgrade the product operation decision-making platform "Liying" - DeInsight, which was built in the early stage, through large-scale models, which greatly reduces the threshold for using product operation decisions.
Based on the technology upgrade of the large model, DeInsight has broken through the limitations of the human brain's understanding of multi-dimensional, multi-indicator, long-term, and large data volumes. Through multiple rounds of intelligent dialogue, it can focus on any SKU according to the needs of business personnel in a few seconds. Perform multi-dimensional and complex calculation and review of massive data, dynamically present massive indicators in one interface in real time, and dynamically play the data analysis results of products like a movie, helping decision makers to read, understand and gain insight into the whole life cycle of products in a very short time The operation process provides intelligent decision support for the ordering, replenishment, distribution, and transfer of goods.
02 Released Wanjie engine with China Nuclear Equipment Institute to explore the intelligent upgrade of key scenarios
In June of this year, Dipu Technology and the China Nuclear Equipment Technology Research Institute jointly established the "Large Model Data Intelligence Laboratory" and released the Wanjie AGI engine. Based on the Wanjie engine, the two parties can provide the CNNC system and end customers with industry text report generation, intelligent question and answer dialogue, industry domain knowledge base and other services, intelligent image recognition and other related intelligent applications, and accelerate the implementation of large models in industrial enterprise-level application fields .
Build a complete data+AI product system to fully empower the digital transformation of enterprises
The future of enterprise data scenarios is model-based general intelligent business applications.
Dipu Technology is committed to the industrial innovation of "Data+AI" in the enterprise B-end market, and has built a complete product system. The intelligent lake warehouse platform FastData has achieved a significant market share in the construction of enterprise data infrastructure. This year, Dipu Technology and its in-depth cooperation customers jointly explored the practice of large-scale models in the field of enterprise services, created a large-scale model landing tool chain platform FastAGI, and actively invested in and deployed the basic computing power and technical capabilities required by Deepexi, a large-scale model in the Dipu Technology industry . , to better serve customers from all walks of life.
On the same day, the conference released the "2023 China AI Commercial Landing Investment Value Research Report" released by EO Think Tank, and Dipu Technology was selected into the list of vertical scene service providers with high investment value in AI commercial landing and the map of the entire artificial intelligence industry chain.