Information retrieval, dialogue method, device, equipment, storage medium and program product

By using a multimodal data rewriting model, data from multiple modalities is rewritten into target text of a single text modality for retrieval. This solves the problem of poor accuracy caused by the reliance on knowledge bases in existing multimodal retrieval technologies, and achieves flexible and accurate information retrieval.

CN122220486APending Publication Date: 2026-06-16ALIBABA (CHINA) CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
ALIBABA (CHINA) CO LTD
Filing Date
2024-12-13
Publication Date
2026-06-16

AI Technical Summary

Technical Problem

When performing information retrieval based on input data with multiple modalities, a large amount of pre-configuration of the knowledge base is required, resulting in poor accuracy of the retrieval results, especially when the knowledge base is not rich.

Method used

By using a multimodal data rewriting model, data from multiple modalities can be rewritten into target text of a single text modality. The target text can then be used for retrieval, reducing reliance on pre-defined knowledge bases and improving the flexibility and accuracy of the retrieval process.

Benefits of technology

It enables the conversion of multimodal retrieval into single-text modal retrieval, reduces reliance on knowledge bases, improves the accuracy and flexibility of information retrieval, meets users' in-depth retrieval needs, and enhances user experience.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122220486A_ABST
    Figure CN122220486A_ABST
Patent Text Reader

Abstract

The application provides an information retrieval method, a dialogue method, an apparatus, a device, a storage medium and a program product. The information retrieval method comprises the following steps: receiving user input information, wherein the user input information comprises data in multiple modalities; performing data modality rewriting on the data in multiple modalities to obtain target text; the target text is used to describe the content represented by the data in multiple modalities; and performing retrieval based on the target text to obtain a target retrieval result. The application can improve the accuracy of the target retrieval result.
Need to check novelty before this filing date? Find Prior Art