Interactive Terminal Voice Topic Segmentation for Visitor Intent

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional methods for interacting with visitors require significant human and material resources, leading to low efficiency and high workload in understanding and responding to visitor intentions.

Innovation Solution

An interactive method and terminal that utilize voice data processing, including word segmentation and topic generation models, to determine visitor topics and select appropriate interactive content based on appearing probabilities, reducing the need for human intervention.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If manual interaction methods are used to understand and respond to visitor intentions, then staff members can directly communicate with visitors, but significant human and material resources are consumed leading to low efficiency and high workload

Engineering Contradiction:
Improveinteraction efficiencyVSAvoidhuman resource consumption
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The system enables automated visitor interaction through voice recognition and natural language processing technologies. The interactive terminal independently analyzes visitor voice data, performs word segmentation, determines topics using topic generation models, and selects appropriate responses without requiring manual staff intervention, thus achieving self-service automation

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent replaces the mechanical manual interaction system with an automated electronic system. Voice recognition technology converts visitor speech into text, which is then processed through computational algorithms including word segmentation and topic modeling, ultimately generating automated responses that substitute for human staff actions

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Productivity

If automated voice recognition and topic generation models are used to determine visitor intentions, then interaction efficiency is improved, but the system complexity increases due to multiple processing steps

Engineering Contradiction:
Improveinteraction efficiencyVSAvoidsystem complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The system segments the complex task of understanding visitor intentions into distinct processing stages: voice data acquisition, word segmentation to extract feature words, topic determination using topic generation models, and response selection based on appearing probabilities. This segmentation makes the complex system more manageable and modular

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces intermediate processing components including word segmentation modules that extract feature words from recognized text, and topic generation models that bridge the gap between raw text and meaningful visitor intentions. These intermediaries facilitate the transformation of unstructured voice data into structured actionable insights

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentUS11069338B2Interactive method, interactive terminal, storage medium, and computer device
Publication Date: 2021.07.20 TENCENT TECHNOLOGY (SHENZHEN) CO LTD
  • US11069338B2 patent drawing
  • US11069338B2 patent drawing
  • US11069338B2 patent drawing

AI summary

An interactive method is provided for an interactive terminal. The method includes obtaining visitor voice data of a visitor; and performing word segmentation on a text obtained by recognizing the visitor voice data, to obtain a feature word set. The method also include determining, according to a topic generation model and the feature word set, a determined topic to which the text belongs; and separately obtaining appearing probabilities corresponding to feature words in the feature word set when the feature words belong to the determined topic. The method further includes selecting a feature word from the feature words of the feature word set according to the appearing probabilities corresponding to the feature words; and obtaining and outputting visitor interactive content corresponding to the selected feature word.