Event-Tracked Conversation Scenario Models for Service Process Guidance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing conversation scenario models struggle to fully cover complex service processes and lack the ability to effectively control the conversation process, particularly when users fail to perform operations according to linguistic representations.
Innovation Solution
A method involving event-tracking to capture user behavior data, processing it into flagged guide information, and integrating this information into question-answer conversation texts for model training, enhancing the model's control capabilities to guide users through the service process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the conversation scenario model is trained only with text information translated from conversation content, then the model can interact with users and guide service processes, but the model cannot fully cover the service process and lacks the ability to control the conversation process when users fail to perform operations according to linguistic representation
Solution Approach 1:
The patent merges two types of information into the training data: (1) text information translated from conversation content, and (2) event-tracking information from information interaction pages. This combination allows the model to learn both linguistic representations and actual user operation sequences, thereby achieving full service process coverage and improving control capability.
Solution Approach 2:
The patent introduces event-tracking information as an intermediary element that bridges the gap between conversation text and actual user operations. This intermediary provides structured operation data that helps the model understand and control the service process flow,弥补ing the deficiency of text-only training data.
2Adaptability or versatility
If more guide nodes are arranged in different orders in the service process, then the service process coverage is improved, but the model complexity increases and becomes difficult to manage
Solution Approach 1:
The patent performs preliminary action by collecting and organizing event-tracking information from information interaction pages before model training. This pre-processing step structures the operation data in advance, allowing the model to learn established operation sequences without requiring complex internal structures to manage multiple guide nodes.
Solution Approach 2:
The patent uses event-tracking information as a copy of actual user operation sequences. Instead of creating complex model structures to represent all possible service process variations, the model learns from copied real operation data, simplifying the model structure while maintaining high service process coverage.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method, an apparatus and a device for generating a conversation scenario model are provided, which relate to the technical field of intelligent conversation and can generate an information interaction model which comprehensively covers a service process so that a robot customer service controlled by the model can guide a user to trigger a text conversation which is more in line with process operation. The method includes: acquiring a question-answer conversation text formed in a historical information interaction process of a conversation scenario, wherein the question-answer conversation text comprises multiple rounds of text interaction information; adding flagged guide information into the question-answer conversation text, wherein the flagged guide information is field information obtained by processing event-tracking records in an information interaction page; and performing model training by using a question-answer conversation text with the flagged guide information to generate an information interaction model of the conversation scenario.