Dialog method and apparatus, and device and medium

By acquiring user characteristics from social networking platforms and using anchor point matching and ranking models to determine the optimal topic copy, the problem of user confusion and low interaction efficiency in private chats was solved, enabling personalized and natural conversation initiation, and improving user experience and platform activity.

WO2026123716A1 Publication Date: 2026-06-18SHANGHAI SOULGATE TECH CO LTD

Patent Information

Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
SHANGHAI SOULGATE TECH CO LTD
Filing Date
2025-07-31
Publication Date
2026-06-18

AI Technical Summary

Technical Problem

Existing social networking platforms lack effective icebreakers, leading to user confusion, insufficient personalization, low interaction efficiency, and poor user experience when chatting privately, thus affecting user satisfaction and platform activity.

Method used

By acquiring user characteristics for anchor point matching, and using a ranking model to determine the optimal topic copy from the copy library as an opening remark, personalized and natural conversation can be initiated.

🎯Benefits of technology

It improved the efficiency and fluency of initiating conversations, enhanced the communication experience between users, and increased user satisfaction and platform activity.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
Patent Text Reader

Abstract

A dialog method and apparatus, and a device and a medium, which relate to the field of data processing. The method comprises: when a first user initiates a chat request for a second user, instantly acquiring user features of both parties; for each anchor point, determining whether a feature among first user features that corresponds to the anchor point matches a feature among second user features that corresponds to the anchor point; if the feature among the first user features that corresponds to the anchor point matches the feature among the second user features that corresponds to the anchor point, determining that anchor point matching is successful, otherwise, determining that the anchor point matching fails; sequentially traversing each anchor point to obtain successfully matched anchor points; after determining the matched anchor points, first performing screening to obtain a plurality of candidate topic copies, then using a sorting model to perform personalized sorting on the candidate copies on the basis of the user features of both parties, so as to determine an optimal topic copy, such that the efficiency of determining a final topic copy is higher; and using the optimal topic copy as an opening line, thereby effectively promoting the smooth development of a dialog between both parties. In this way, the startup efficiency of dialogs is improved, making the communication between users more natural and fluent.
Need to check novelty before this filing date? Find Prior Art