Multi-way Tree Response Forest for Multi-domain Dialogue
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current human-machine dialogue products are limited by a single-domain and single-role implementation mode, resulting in undesirable applicability and response broadcasting effects.
Innovation Solution
A method and apparatus using a multi-way tree data structure with at least three layers of nodes (root, domain, and role nodes) to generate responses, allowing for semantic parsing and slot replacement to create diverse and compliant responses across multiple domains and roles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a single-domain and single-role implementation mode is used, then the product structure is simple, but the applicability is poor and response broadcasting effect is undesirable
Solution Approach 1:
The patent segments the response generation system into multiple independent domains and roles. Each domain (e.g., weather, news, sports) and each role (e.g., news anchor, weather reporter) is treated as a separate module with its own response templates and broadcasting characteristics. This segmentation allows the system to handle multiple domains and roles while maintaining manageable complexity through modular architecture.
Solution Approach 2:
The patent creates a universal response generation framework that can serve multiple domains and roles simultaneously. The system uses a common technical infrastructure (semantic parsing, template matching, slot replacement) that works across different domains and roles, enabling one system to perform multiple functions. Each domain and role configuration can be independently adjusted while sharing the core processing mechanisms.
2Adaptability or versatility
If a single-domain and single-role implementation mode is used, then the system is easy to operate, but the response broadcasting effect is undesirable
Solution Approach 1:
The patent performs preliminary actions by pre-defining domain configurations, role configurations, and response templates for each domain-role combination. The system pre-processes and stores structured response templates with placeholders for various slots, and pre-configures the semantic parsing rules for different domains. This preliminary preparation reduces operational complexity during actual use, as the system can directly match and fill templates without requiring complex real-time decisions.
3Adaptability or versatility
If multi-domain and multi-role response generation is implemented, then applicability and response broadcasting effect are improved, but the system complexity increases
Solution Approach 1:
The patent implements a nested structure where domains contain multiple roles, and each domain-role combination contains specific response templates. The system hierarchy is organized as: System → Multiple Domains → Multiple Roles per Domain → Multiple Response Templates per Role. This nesting allows the system to manage complexity by organizing elements at different levels, where each level can be independently configured and maintained.
Solution Approach 2:
The patent manages multi-domain multi-role complexity by parameterizing the system configuration. Each domain and role is defined by a set of parameters (domain type, role type, template identifiers, slot definitions). The system dynamically selects and adjusts these parameters based on the input context, allowing flexible configuration without hardcoding complex conditional logic. This parameter-driven approach simplifies the management of multiple domains and roles.
Data Source
AI summary
A method and apparatus for generating a response based on artificial intelligence, and a storage medium. The method comprises: generating a response forest which employs a multi-way tree data structure, the multi-way tree at least comprising three layers of nodes: a root node, domain nodes and role nodes arranged in a top-bottom order, each leaf node respectively corresponding to at least one response template corresponding to information on the leaf node path; obtaining a user question, searching the response forest according to the user question to obtain a leaf node corresponding to the user question, and regarding a response template corresponding to the obtained leaf node as a candidate response template; generating a to-be-broadcast response according to the user question and the candidate response template, and broadcasting the to-be-broadcast response to the user. The solution of the present disclosure exhibits wide applicability and can improve the response broadcasting effect.


