Cloud Semantic Indexing Platform for Web Objects
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional techniques for processing web objects lack efficient methods for generating and associating semantic information, which hampers the performance of search engines and data analytics engines, as they often require separate semantic analysis for each web object.
Innovation Solution
A cloud-based web object indexing platform utilizing machine learning engines, graph analyzers, and semantic information generators to process web objects, such as text, images, and video, and generate RDF triples for semantic information, which is then incorporated into metadata, reducing the need for separate analysis by search and data analytics engines.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If separate semantic analysis is performed for each web object by search engines and data analytics engines, then semantic information can be obtained, but processing time and computational resources are significantly consumed
Solution Approach 1:
The patent implements preliminary semantic analysis by creating a cloud-based platform that pre-processes web objects and generates semantic information before search or analytics operations occur. The system includes web object indexing applications that perform semantic analysis, machine learning engine training, and graph analytics in advance, storing results in data lakes for rapid retrieval during actual search operations, thereby eliminating the need for repeated semantic analysis on each query
Solution Approach 2:
The patent introduces an intermediary cloud-based semantic indexing platform between web object storage and search/analytics engines. This platform includes indexing applications that act as mediators, performing semantic analysis and generating structured semantic information that is then made available to multiple search and analytics engines simultaneously, reducing redundant processing across different systems
2Productivity
If conventional processing techniques are used for web objects, then system complexity is maintained, but semantic information generation and association efficiency is insufficient
Solution Approach 1:
The patent implements a universal cloud-based platform that performs multiple functions: web object indexing, semantic analysis, machine learning model training, graph analytics, and result storage. The indexing applications can process different types of web objects (text, images, video) and generate various forms of semantic information, making the system multi-functional and reducing the need for separate specialized systems
Solution Approach 2:
The patent introduces an intermediary cloud-based semantic indexing platform between web object storage and search/analytics engines. This platform includes indexing applications that act as mediators, performing semantic analysis and generating structured semantic information that is then made available to multiple search and analytics engines simultaneously, reducing redundant processing across different systems
3Measurement precision
If machine learning engines are trained and graph analytics are performed to generate semantic information, then search and analytics quality is improved, but computational resources and processing time are increased
Solution Approach 1:
The patent implements preliminary semantic analysis by creating a cloud-based platform that pre-processes web objects and generates semantic information before search or analytics operations occur. The system includes web object indexing applications that perform semantic analysis, machine learning engine training, and graph analytics in advance, storing results in data lakes for rapid retrieval during actual search operations, thereby eliminating the need for repeated semantic analysis on each query
Solution Approach 2:
The patent implements continuous learning and optimization by training machine learning engines on newly ingested data and updating semantic models continuously. The system performs ongoing graph analytics and semantic enrichment, maintaining up-to-date semantic information in data lakes, which provides continuous value to search and analytics operations without requiring repeated full-processing cycles
Data Source
AI summary
An apparatus in one embodiment comprises a cloud-based web object indexing platform configured to communicate with multiple web servers over at least one network. The cloud-based web object indexing platform comprises a plurality of indexing applications having respective machine learning engines, with a given one of the indexing applications being configured to receive web objects over the network from one or more of the web servers and to process the received web objects utilizing its corresponding machine learning engine. The given indexing application is further configured to generate semantic information for association with particular ones of the web objects based at least in part on processing results provided by the corresponding machine learning engine. The semantic information is made accessible by the cloud-based web object indexing platform in association with the particular web objects over the network, for example, to a search engine or data analytics engine.