Semantic Search Interface for Intent-Based Data Visualization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current search systems for data repositories are limited in expressivity, relying on keyword-based methods that fail to support structured query types and do not effectively utilize the sparse textual information within data sources and visualizations, restricting users to specific responses and lacking the ability to interpret user intent beyond simple keyword matching.
Innovation Solution
A hybrid semantic search approach that combines natural language question and answering interfaces with data repositories, utilizing entity recognition, word disambiguation, and relationship extraction to interpret user intent, generating visualization responses and pre-authored charts, and supporting faceted search to refine results based on parameters like author name, time, and chart type.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If keyword-based search methods are used, then content retrieval is supported, but structured query types and contextual understanding are limited
Solution Approach 1:
The patent introduces natural language processing and semantic analysis as intermediary layers between the user's keyword input and the data repository. These intermediaries interpret the intent behind keywords, perform entity recognition, and generate structured queries, thereby preserving contextual information without requiring the user to directly construct complex query structures.
Solution Approach 2:
The patent replaces traditional mechanical keyword-matching search mechanisms with intelligent text processing systems that use natural language understanding, entity recognition, and semantic analysis. This substitution enables the system to comprehend contextual meaning and generate appropriate structured queries automatically.
2Adaptability or versatility
If natural language question and answering interfaces are used, then fact-finding inquiry is supported, but content and document discovery are limited
Solution Approach 1:
The patent merges the strengths of natural language question-answering interfaces with keyword-based search capabilities into a unified hybrid search system. This combination allows the system to handle both fact-finding inquiries through natural language processing and content discovery through enhanced keyword matching, thereby supporting diverse query types while maintaining document retrieval effectiveness.
3Ease of operation
If current search systems are used, then simple keyword search is supported, but user expressivity and search refinement are limited
Solution Approach 1:
The patent implements a dynamic search system that adapts to user needs by providing multiple search modes and refinement options. The system can dynamically switch between simple keyword matching and complex semantic analysis based on the query type, and offers faceted search capabilities that allow users to refine results interactively, thereby maintaining ease of operation while enhancing search expressivity.
4Quantity of substance
If data repositories have sparse textual information, then data storage efficiency is improved, but searchability and content discovery are reduced
Solution Approach 1:
The patent replaces traditional text-based search mechanisms with intelligent text processing systems that can extract meaningful information from sparse textual data. Through entity recognition, relationship extraction, and semantic analysis, the system generates structured representations of the data that enhance searchability without requiring additional textual information, thereby maintaining storage efficiency while improving content discoverability.
Data Source
AI summary
A method provides visual analysis of datasets. A system receives a natural language search query that is directed to data repositories including data sources and data visualizations. The system parses search tokens to determine if the natural language search query contains analytic intents. The system also determines if the search tokens match fields in one or more data sources, using a semantic search. When (i) the search tokens match fields in the one or more data sources and (ii) the natural language search query contains analytic intents, the system generates and displays visualization responses. When (i) the search tokens do not match fields in the data sources or (ii) the natural language search query does not contain the analytic intents, the system displays pre-authored content from the data visualizations.