Compressed Word Vector Tree for Local File Search Expansion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current file searching technologies are limited by the number of designated words that can be used to search for files, restricting the retrieval of files based on associated metadata, such as an image file only being searchable by its explicitly designated word without related terms.
Innovation Solution
The system generates a compressed word vector structure and a tree look-up structure to expand the number of searchable words by determining additional words related to the query through proximity analysis, allowing files to be searched using a broader range of terms without network access, while preserving storage space and reducing processing time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If files are searched using only explicitly designated words, then storage space is preserved and processing time is reduced, but the number of searchable terms is limited and search functionality is restricted
Solution Approach 1:
The patent segments the large vocabulary into multiple clusters using a tree structure, where each node represents a cluster of semantically related words. This segmentation allows the system to handle large numbers of search terms without overwhelming memory resources, as only relevant clusters need to be loaded into memory during search operations.
Solution Approach 2:
The patent transforms the traditional flat word-list approach into a multi-dimensional hierarchical tree structure. Words are organized across multiple levels of abstraction, with each level representing a different dimension of semantic organization. This dimensional transformation enables efficient navigation and retrieval without requiring all words to reside in memory simultaneously.
2Reliability
If a compressed word vector structure is implemented locally, then search capabilities are expanded without network access, but storage space requirements increase
Solution Approach 1:
The patent implements local quality by storing compressed word vectors only at specific nodes within the tree structure rather than uniformly across all nodes. This allows the system to optimize storage by placing computational resources where they are most needed, maintaining search functionality while minimizing overall storage requirements.
Solution Approach 2:
The patent applies partial action by loading only the necessary portions of the word vector structure into memory during search operations, rather than loading the entire structure. This selective loading approach maintains search availability while keeping storage and memory usage at minimal necessary levels.
3Adaptability or versatility
If additional related words are determined through proximity analysis, then the number of searchable words is expanded, but processing time increases
Solution Approach 1:
The patent performs preliminary action by pre-computing and storing word vectors and their semantic relationships in the tree structure during an offline indexing phase. This pre-processing allows the system to quickly retrieve related words during online search operations without performing computationally intensive proximity analysis in real-time, thus minimizing processing time during actual searches.
Data Source
AI summary
A device implementing a system for expanded search includes a processor configured to identify plural words, and generate, for each word of the plural words, a word vector based on a proximity of the word relative to other words of the plural words, the word vector comprising plural dimensions. The processor is further configured to create a compressed word vector structure comprising clusters of subsets of the plural dimensions across the word vectors, each cluster including similar values of the respective dimensions, convert the word vectors to points on at least one plane, and partition the at least one plane into nested groupings of the points based on a threshold number of points per nested grouping. The processor is further configured to create a tree look-up structure of the nested groupings, and provide the compressed word vector structure and the tree look-up structure to a client device.