Distributed Clustering Agents for Textual Information
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current information retrieval systems require centralized repositories for document updates, leading to significant processing, communication, and storage demands when documents are added or removed, which is inefficient in distributed systems.
Innovation Solution
A distributed clustering method that operates without a centralized repository, using a hierarchical system of program modules across multiple computers to determine document similarity and route new documents for storage, allowing for efficient use of storage capacity and system expansion.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a centralized repository is used to maintain document information for clustering, then document updates can be centralized and managed, but significant processing, communication, and storage capacity is required when documents are added or removed
Solution Approach 1:
The patent divides the centralized clustering system into distributed cluster agents operating on multiple computers. Each agent maintains local cluster information independently, eliminating the need for a single centralized repository. When documents are added or removed, only local cluster agents need to process updates, significantly reducing the processing and storage capacity requirements compared to a centralized system.
Solution Approach 2:
The patent transitions from a single-dimensional centralized storage model to a multi-dimensional distributed architecture where cluster information is replicated across multiple computers and networks. This dimensional expansion allows the system to handle document updates in parallel across multiple nodes, reducing the burden on any single system and optimizing overall processing capacity.
2Power
If distributed clustering is implemented without centralized repository, then processing and storage demands are reduced, but system complexity increases due to multiple program modules across computers
Solution Approach 1:
The patent designs cluster agents with universal functionality that can operate independently on any computer in the distributed network. Each agent performs the same core functions of maintaining cluster information, comparing documents, and routing updates. This multi-functionality reduces system complexity by using standardized, interchangeable components rather than specialized modules for each function.
Solution Approach 2:
The patent implements feedback mechanisms where cluster agents continuously exchange information about document clusters and routing decisions. This feedback loop allows the distributed system to self-regulate and maintain consistency without centralized control, reducing the need for complex coordination protocols and simplifying the overall system architecture.
3Productivity
If information agents are used to collect and filter information, then information retrieval capability is improved, but the system requires significant communication and processing resources
Solution Approach 1:
The patent extracts the information filtering and clustering functions from a centralized processing model and distributes them to local cluster agents. Each agent filters and clusters information locally based on its assigned topic area, eliminating the need for all information to be communicated to a central system. This extraction of functions to the edge of the network significantly reduces communication and processing resources while maintaining retrieval capability.
Solution Approach 2:
The patent implements local quality by having each cluster agent specialize in specific topic areas and maintain optimized cluster information for those topics. This localization allows agents to process and filter information more efficiently for their specific domain, reducing the overall communication overhead while improving the quality and relevance of information retrieval for different subject areas.
Data Source
AI summary
A computer method and system for storing, retrieving and displaying information has a multiplexing agent (20) that calculates a new document vector (25) for a new document (21) to be added to the system and transmits the new document vector (25) to master cluster agents (22) and cluster agents (23) for evaluation. These agents (22, 23) perform the evaluation and return values upstream to the multiplexing agent (20) based on the similarity of the document to documents stored under their control. The multiplexing agent (20) then sends the document (21) and the document vector (25) to the master cluster agent (22), which then forwards it to a cluster agent (23) or creates a new cluster agent (23) to manage the document (21). The system also searches for stored documents according to a search query having at least one term and identifying the documents found in the search, and displays the documents in a clustering display (80) of similarity so as to indicate similarity of the documents to each other.