Expert Conversation Builder Using Directed Acyclic Graphs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing conversation-based customer support systems lack seamless integration of user and authoring modes, struggle with handling multiple answers, and fail to facilitate changes in previously submitted answers, making them inefficient for robust conversational diagnostics.
Innovation Solution
A system utilizing a directed acyclic graph (DAG) to represent questions and responses, allowing for dynamic traversal and modification of conversation paths, enabling collaborative authoring and real-time interaction with users, and supporting multiple answers through a web-based dialogue authoring tool.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a tree-based hierarchy is used to organize diagnostic information, then the system structure is simple and easy to implement, but the system cannot handle multiple answers or allow changes to previously submitted answers
Solution Approach 1:
The patent segments the rigid tree-based hierarchy into a more flexible graph structure where nodes represent questions and edges represent possible answers. This segmentation allows multiple paths and answers while maintaining manageable system organization through modular question-node units that can be independently authored and modified.
Solution Approach 2:
The system transitions from a static tree structure to a dynamic graph-based conversation model where the diagnostic path can adapt in real-time. Users can change previously submitted answers, and the system dynamically recalculates the diagnostic path based on new input, enabling flexible adaptation without re-authoring the entire diagnostic flow.
2Reliability
If specialized authoring tools are used to create robust conversational diagnostic tools, then the diagnostic capability is improved, but the integration with user experience is poor and switching between user and authoring modes is difficult
Solution Approach 1:
The web-based authoring tool is designed to be universal, serving both as a powerful diagnostic authoring environment and as the user interface for end-users. The same interface that allows SMEs to create and modify diagnostic conversations also enables users to interact with the system, eliminating the need for separate user and authoring modes and allowing seamless switching between roles.
Solution Approach 2:
The patent merges the authoring functionality and user interaction into a single unified web-based interface. The authoring tool combines diagnostic creation capabilities with real-time user interaction features, allowing SMEs to author, test, and refine diagnostic conversations within the same environment that users will experience, thereby integrating authoring and user experience.
3Measurement precision
If all data in the corpus is searched entirely at one time, then comprehensive diagnosis is achieved, but the system is inefficient and requires returning to the start to handle answer changes
Solution Approach 1:
The system performs preliminary actions by pre-organizing diagnostic information into a graph structure with questions as nodes and answers as edges before user interaction. This pre-computation of the diagnostic graph allows the system to efficiently navigate and search through diagnostic paths during user conversations without needing to reprocess the entire corpus, enabling both comprehensive diagnosis and efficient handling of answer changes.
Solution Approach 2:
The patent replaces the mechanical approach of searching the entire corpus linearly with a graph-based navigation system. Instead of dumping and searching all data at once, the system uses the structured graph to efficiently traverse relevant diagnostic paths based on user responses, substituting brute-force searching with intelligent graph traversal that maintains comprehensiveness while improving productivity.
4Ease of manufacture
If a rigid conversation structure is used, then the system is easy to author and maintain, but users cannot change previously submitted answers or discuss specific points with other users and SMEs
Solution Approach 1:
The system implements dynamic conversation structures where the diagnostic graph can adapt during user interactions. Users can modify previously submitted answers, and the system dynamically adjusts the conversation path based on new input. The graph structure naturally supports these changes by allowing alternative paths and enabling users to revisit earlier nodes in the diagnostic flow.
Data Source
AI summary
An expert conversation builder contains a knowledge database that includes a plurality of dialogues having nodes and edges arranged as directed acyclic graphs. Users and authors of the system interface with the knowledge database through a graphical interface to author dialogues and to create expert conversations as threads traversing the node in the dialogues.