System and method for realizing multi-round deep retrieval on television based on TmallGenie

Abstract

The invention belongs to the technical field of multi-round depth retrieval, and particularly to a system and a method for realizing multi-round deep retrieval on a television based on TmallGenie. Thesystem comprises a TmallGenie voice acquisition module, a TmallGenie skill platform module and an account login authorization binding module. Multi-round depth retrieval is executed after a voice text is received. Deep learning model training is performed on the preprocessed intention and entity data set to obtain a model; the voice text is transmitted to the model, the intention of the user is predicted, the retrieval request of the user in the last round needs to be subjected to associated logic judgment and multi-round retrieval logic, multi-round intention and entity integration can be carried out according to the associated information of the retrieval request in the previous round and the retrieval request in the next round, and therefore a retrieval engine is called to push a multi-round deep retrieval result to a television to be displayed; according to the system and the method, through voice interaction with the TmallGenie, multiple rounds of dialogue type deep retrieval canbe quickly completed, so that resources wanted in the television can be accurately retrieved.

Description

technical field [0001] The invention relates to the technical field of multiple rounds of in-depth retrieval, in particular to a system and method for realizing multiple rounds of in-depth retrieval on TV based on Tmall Genie. Background technique [0002] With the widespread popularization of the Internet and the further development of 3C integration, Internet TV has become the mainstream of the market. At the same time, with the diversity and richness of Internet resources, the popularity of intelligent voice terminal equipment, and the inconvenience of retrieving resources on TV. Therefore, a method is proposed, based on the connection between Tmall Genie and smart TV, to realize the function of using Tmall Genie's voice to perform multiple rounds of in-depth retrieval of resources on TV. [0003] Defects of existing technology: [0004] 1) Difficulty in retrieval: The current retrieval interface in the smart TV system is based on using the remote control to input the f...

AI Technical Summary

Problems solved by technology

[0006] 3) The search depth is single: the current mainstream search is based on keywords and words, and it is impossible to achieve multiple rounds of in-depth understanding of intent searches. For example: the first round of search requests is that I want to watch the movie of Deng Chao’s wife, and the second round of search requests is for comedy

In this scenario, there is no way to retrieve the comedy movies of Sun Li (Deng Chao’s wife) based on the combination of the two user intentions and understanding in a single round of keyword retrieval.

Images