Contextual Data Browsing via Hierarchical Sentence Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In enterprise environments, comprehending the complex web of relationships between hundreds of thousands of data objects stored across different data sources is challenging due to the difficulty in presenting these relationships in a human-readable and easily understandable format.
Innovation Solution
A system and method that generate human-readable sentences representing relationships between data objects, organized in a hierarchical format for display, allowing users to browse and search data objects by understanding the context through subject-verb-object sentences, with attributes of selected data objects provided for additional details.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If data objects and relationships are collected to build a single datastore, then the completeness of information is improved, but the complexity of presentation and human comprehension deteriorates
Solution Approach 1:
The patent segments the complex web of data relationships into hierarchical groups and categories. Data objects are organized into parent-child relationships with multiple levels of grouping, allowing the complete information to be divided into manageable segments that can be progressively explored without overwhelming the user.
Solution Approach 2:
The patent introduces a hierarchical dimension to organize data relationships. Instead of presenting all relationships in a flat structure, data objects are arranged across multiple hierarchical levels with parent-child relationships, adding a vertical dimension to the data presentation that makes complex relationships more comprehensible.
2Loss of information
If all data objects are displayed to provide complete information, then the completeness of data is improved, but the ease of operation and user comprehension deteriorates
Solution Approach 1:
The interface segments data objects into hierarchical groups that can be progressively explored. Users start with high-level groupings and can drill down into sub-groups as needed, allowing complete data to be accessed in manageable portions rather than all at once.
Solution Approach 2:
The patent implements dynamic interaction where users can select and explore data objects at different hierarchical levels. The system responds to user selections by displaying relevant sub-groups and details, making the data exploration process adaptive to user needs rather than static and overwhelming.
3Ease of operation
If hierarchical organization of sentences is implemented, then the ease of comprehension is improved, but the device complexity increases
Solution Approach 1:
The patent implements a universal hierarchical structure that can accommodate different types of data objects and relationships through a common framework. The same parent-child grouping mechanism and sentence generation approach work across diverse data types, reducing the need for separate handling logic for different data categories.
Solution Approach 2:
The system uses template-based sentence generation where standardized sentence patterns are copied and adapted for different data objects. This allows complex hierarchical relationships to be represented through repeated application of simple sentence templates, reducing the complexity of generating diverse representations from complex data structures.
Data Source
AI summary
Systems and methods for searching and browsing a data store of contextually related data objects. The system includes a search/browse module that receives a search query. The search/browse module identifies data objects that match the search query and generates sentences from data objects that are contextually related to the matching data objects. The sentences are human-readable sentences, for example in subject-verb-object format, where each sentence represents the relationship between two data objects. The sentences are output for display as a hierarchy of sentences. Additionally, a user can browse the data store of contextually related data objects by selecting a sentence that is displayed to the user. The search/browse module then outputs attributes of the data object represented by the sentence for display in two separate regions of a user interface.