Semantic Search Index for Portable Terminals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data searching methods in portable terminals are inefficient when searching for various types of data, as they rely on exact keyword indices, which are not known for the desired content, limiting the ability to find relevant information effectively.
Innovation Solution
A semantic-based searching apparatus and method that uses a storage unit with an associative search structure, including feature metadata, semantic entity metadata, and semantic relation metadata indices, allowing for cross-referencing and generating search results based on input queries, even when the exact name of the object is unknown, with additional metadata provided by a semantic metadata server.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If keyword-based search indices are used, then search speed for exact matches is improved, but search capability for unknown or partial keywords deteriorates
Solution Approach 1:
The patent introduces semantic metadata as an intermediary layer between the search query and the stored data indices. This semantic metadata includes semantic entities and semantic relations that mediate the search process, allowing the system to understand the meaning behind queries and retrieve relevant information even when exact keywords are unknown. The semantic metadata acts as a bridge that translates user queries into meaningful search results across multiple data types.
Solution Approach 2:
The patent extends the traditional keyword-based search by adding a semantic dimension. Instead of only searching through literal keywords, the system now operates in a multi-dimensional search space that includes semantic entities, semantic relations, and feature metadata. This dimensional expansion allows the search to function effectively in scenarios where exact keyword matching fails, such as when users describe objects indirectly or when dealing with diverse data types.
2Device complexity
If traditional keyword indices are embedded in portable terminals, then device simplicity is maintained, but search effectiveness for various data types deteriorates
Solution Approach 1:
The patent creates a universal search mechanism that can handle multiple data types (text, image, audio, video) through a single semantic-based approach. The semantic metadata structure serves as a universal index that can be applied across different data formats, allowing the portable terminal to maintain a unified, simple architecture while achieving enhanced search effectiveness. The same semantic indexing mechanism works for searching contacts, messages, media files, and other data types.
Solution Approach 2:
The system performs preliminary semantic analysis and metadata generation when data is stored or updated. By pre-processing data to extract semantic entities, relations, and features before the actual search occurs, the system prepares the data in advance for efficient retrieval. This preliminary action reduces the complexity of the search operation itself, allowing fast and effective searching across various data types without requiring complex real-time processing during the search query.
3Measurement precision
If exact keyword matching is required, then search precision for known terms is improved, but user convenience when unsure of exact terms deteriorates
Solution Approach 1:
The semantic-based search system provides feedback in the form of relevant search results that reflect the user's intended meaning, even when the query uses imprecise or partial keywords. The system analyzes the semantic content of the query and returns results based on semantic similarity and relations, effectively communicating back to the user what information is most relevant to their query intent. This feedback mechanism helps users refine their search understanding and find accurate results without needing to know the exact terminology.
Solution Approach 2:
The patent changes the search parameter from literal keyword matching to semantic meaning matching. Instead of requiring exact keyword correspondence, the system uses semantic entities and relations as the matching criteria. This parameter change allows the search to be more flexible and user-friendly, as it can match queries based on their semantic content rather than requiring precise keyword usage. The system maintains precision by using semantic relations to ensure relevant results while improving ease of operation through keyword flexibility.
Data Source
AI summary
A semantic-based searching apparatus and method which can search for desired information from various types of media using associative properties of search target objects are provided. The semantic-based searching apparatus stores an associative search structure that previously stores a semantic index. The associative search structure is to obtain a final search object corresponding to an input search query. The semantic index configuration includes feature metadata used to identify a specific object, semantic entity metadata to indicate semantic entities corresponding to the feature metadata, and semantic relation metadata to indicate a relation between the semantic entities. The semantic-based searching apparatus uses semantic metadata stored in conformity with the semantic index configuration, the semantic index, and the associative search structure to generate a search result corresponding to the input search query.


