Ontology-Driven Chatbot for Complex Digital Content Retrieval
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current information retrieval systems, including search engines and chatbots, struggle to effectively handle open-ended questions and multiple-variable requests, leading to inefficient information retrieval, especially in complex scenarios like enterprise portals.
Innovation Solution
A digital content generation system comprising a content management system, an ontology manager, and a chatbot that synchronizes tags and facets to drive conversations, generating and transmitting relevant digital assets based on user interactions, using a hierarchical ontology and sequence numbers to guide the conversation and retrieve associated assets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional search engines and chatbots are used for information retrieval, then simple keyword searches can be performed, but they fail to effectively handle open-ended questions and multiple-variable requests
Solution Approach 1:
The patent segments complex information retrieval into distinct phases: ontology-based query understanding, dialog state tracking, and context-aware response generation. The system divides the conversation into manageable states (inform, request, confirm) and handles each phase with specialized processing, enabling reliable handling of complex multi-variable queries that traditional single-pass search engines cannot manage
Solution Approach 2:
The patent introduces an ontology manager as an intermediary layer between the user's natural language query and the information retrieval system. This ontology manager translates open-ended questions into structured queries using predefined facets and relationships, serving as a mediator that bridges the gap between unstructured user input and the structured data models required for reliable information retrieval
2Ease of operation
If hierarchical category directories are used for information retrieval, then users can browse organized content, but users lack the expertise and fortitude to navigate the ever-growing hierarchical category directories
Solution Approach 1:
The patent implements a dialog state tracking system that automatically manages the complexity of information retrieval without requiring user expertise. The system self-services by tracking conversation state, identifying user intent, and navigating the information structure autonomously based on natural language cues, eliminating the need for users to manually navigate complex hierarchies while maintaining ease of operation
Solution Approach 2:
The patent transforms the static hierarchical directory into a dynamic conversational interface. The system adapts its navigation path in real-time based on user responses, dynamically adjusting the information retrieval strategy according to the evolving dialog state. This dynamic approach maintains ease of operation by allowing users to speak naturally while the system handles the complexity of navigating structured information spaces
3Productivity
If Ask-Tell search model is used, then users can search for information by keywords, but the model collapses when user's request depends on multiple variables
Solution Approach 1:
The patent maintains continuous information retrieval across multiple conversation turns by tracking dialog state and accumulating context. Instead of treating each query as an independent search operation, the system continuously refines its information retrieval strategy based on the evolving conversation, preserving productivity by avoiding redundant searches while enhancing adaptability through progressive clarification of multi-variable requests
Solution Approach 2:
The patent performs preliminary ontology-based query analysis and dialog state assessment before executing information retrieval operations. By pre-processing the query to identify relevant facets, variables, and expected information types, the system prepares the retrieval strategy in advance, enabling efficient handling of multi-variable requests while maintaining fast response times through pre-computed ontology relationships and contextual predictions
Data Source
AI summary
A system for digital content that includes a content management system, an ontology manager and a chat bot, all executing and in communications coupling on a digital data processing system. The content manager stores a plurality of tagged digital assets. The ontology manager stores a list of (or otherwise maintains) plural facets, each corresponding to one or more tags of the content management system. One or more dialog segments and sequence identifiers are maintained in the ontology manager, as well, each associated with one or more other facets. The chat bot drives a conversation with an end-user based on facets identified as associated with assets in the content management system and using dialog segments associated with those facets, while excluding those facets not so identified. The digital data processing system generates and transmits to the user digital assets identified through that conversation.