As an “accelerator” for enterprise development, scientific and technological innovation research and intellectual property reserves can only promote high-quality enterprise development by actively carrying out scientific and technological innovation research and intellectual property reserves. Recently, there have been frequent reports of Xiaoi Robotics in patent technology innovation, and has obtained many important national invention patents with high gold content. Behind the continuous expansion of the patent bank is Xiaoi Robotics's unremitting technological innovation efforts.
Xiaoi Robot's patent “An Intelligent Customer Service Dialogue Response Generation Method, Device, and Electronic Device” obtained an invention patent certificate, patent number: ZL202010457142.6.
The invention provides an intelligent customer service conversation response generation method, device, and electronic device. The method includes: obtaining this round of user input sessions; coding the current round of user input sessions to obtain current round input vectors; obtaining historical conversation records, which include at least one round of historical user sessions and corresponding historical customer service responses; coding the historical conversation records to obtain hidden historical record vectors; calculating key information vectors corresponding to each preset word slot based on the hidden historical record vectors; calculating correlation vectors between each preset word slot and other word slots based on the key information vector; decoding and obtaining multiple round conversation states according to the correlation vector; The round input vector and the multi-round conversation state described to obtain a response from this round of customer service. Through the above steps, the invention can effectively improve the effectiveness of conversation tracking, thereby improving the performance of intelligent customer service dialogue systems, and has a wide range of application value.
Intercept...

A year ago, ChatGPT came out of nowhere, and the wave of artificial intelligence based on the big model was sparked. By processing language and text, the big model became a “golden key” that opened the door to a new world.
However, this world is not a “sea of words” where words are carefully scrutinized. Especially in the Internet world, most data is images, video, and audio. In order to better describe and express, and change the world, a multi-modal big language model came into being.
According to CCTV, Professor Deng Weihong of Beijing University of Posts and Telecommunications said, “Multiple modes allow human perception and communication to be converted between any information modes.” The multi-modal large model can not only understand multi-modal content, but also generate high-quality content in multi-modal ways such as images, audio, video, and code.
The development of the big model of artificial intelligence in the multi-modal field has brought “disruptive changes” to thousands of industries, creating a new model of human-computer interaction, covering various industries such as intelligent manufacturing, education, finance, media, and healthcare. For example, by generating codes and combining manufacturing processes such as product research and development and process design, production efficiency is fully improved, and the advantages of data as a new factor of production are exploited.
The Huazang General Model can achieve intelligent semantic understanding and knowledge map construction with strong multi-modal capabilities such as literary maps, text retouching, and graphic literature, and has become a capability foundation for the Huazang ecosystem.
The Huazang General Model not only has strong language understanding and generation capabilities, but can also implement various modes such as language, images, and audio...