Ant Group publishes financial model products: financial assistant "Zhixiaobao 2.0" and business assistant "supporting small help", which will be put online after filing.

Thank you, Mr. Air, a netizen of CTOnews.com, for your clue delivery! CTOnews.com Sept. 8 news, a few days ago in Shanghai, the second Bund Congress, Ant Group officially released its financial model.

According to Ant Group, the ant financial model is based on its self-developed basic model, and is deeply customized for the financial industry, and the underlying computing cluster reaches the scale of Vanka. At present, the big model has been fully tested in the Ant Group Wealth and Insurance platform.

At the same time, two products based on the model, "Zhixiaobao 2.0", an intelligent financial assistant, and "Zhixiaosui", an intelligent business assistant serving financial industry experts, have also been officially unveiled. According to reports, two large model products show the full stack layout and progress of ants from basic model to industry model and industrial application. The progress of CTOnews.com with two products is as follows:

"Zhixiaobao 2.0" has begun internal testing for nearly half a year and will be launched after the completion of the relevant filing work.

"Zhixiaozhu" is working with the Ant platform for internal testing and construction, and will create full-chain AI business assistants for financial consultants, insurance agents, investment and research, financial marketing, insurance claims and other financial experts.

Lu Xin, head of Zhixiaobao technology, said that thanks to the improvement of knowledge and expertise brought about by the big financial model, Zhixiaobao 2.0's IQ and financial quotient have been "raised to a new level." it can help users interpret market information in depth, and provide personalized configuration strategies in combination with users' financial goals and investment preferences.

According to Wang Xiaohang, vice president of Ant Group and head of financial models, based on a large number of practices in financial scenarios, ant financial models have formed an architecture driven by "big model + knowledge + service". This architecture has been tested internally on the financial intelligence scene within the ant.

In addition, the large model of Ant Finance claims to extract a total of 600000 + high-quality instruction data from 300 + real industry scenarios, and adopts the combination of intention identification and factual checking. The compliance, security and authenticity of the generated content can be "effectively improved".

Related readings:

"Ant Group will release a full stack layout, pure self-developed AI model tomorrow."