Classification of digital content using aggregate scoring by a grid load balancer configured to assign content to different content analysis engines based on content identifications of a semantics processor
Aggregate scoring is used to help classify digital content such as content uploaded to multi-user websites (e.g., social networking websites). In one embodiment, specific categories are used that relate to a social implication of content. For example, text, images, audio or other data formats can provide communication perceived to fall into categories such as violent, abusive, rights management, pornographic or other types of communication. The categories are used to provide a raw score to items in various groupings of a site's content. Where items are related to other items such as by organizational, social, legal, data-driven, design methods, or by other principles or definitions, the related items' raw scores are aggregated to achieve a score for a particular grouping of items that reflects, at least in part, scores from two or more of the related items.
1. A computer system apparatus for classifying digital content using aggregate scoring comprising:
a computer processor coupled to a memory wherein the computer processor is programmed to classify digital content using aggregate scoring by:
a semantics parser configured to filter content received at web page servers of a multi-user website based on selected criteria to preliminarily identify content for specific analysis;
a plurality of content analysis engines, each configured to perform a specific type of analysis of supplied content and to output a content analysis result;
a grid load balancer, coupled to receive content for specific analysis from the semantics parser and coupled to the plurality of content analysis engines, with the grid load balancer configured to assign content to different content analysis engines based on content identifications of the semantics parser; and
a result aggregator, coupled to the plurality of content analysis engines, with the result aggregator configured to assign raw scores to content analysis results output from each content analysis engine, and to aggregate the raw scores into an aggregate score for a particular grouping of content analysis results that reflects, at least in part, scores from two or more content analysis results.
2. The computer system apparatus of claim 1 where the plurality of content analysis engines includes an early detection content analysis engine.
3. The computer system apparatus of claim 1 where the plurality of content analysis engines includes an image processing content analysis engine.
4. The computer system apparatus of claim 1 where the plurality of content analysis engines includes a test processing content analysis engine.
5. The computer system apparatus of claim 1 where the plurality of content analysis engines includes a video processing content analysis engine.
6. The computer system apparatus of claim 1 where the plurality of content analysis engines includes a copyright processing content analysis engine.
7. The computer system apparatus of claim 1 further comprising:
a comfort zone definition component, coupled to the result aggregator, configured to specify a tolerance level for a classification.
8. The computer system apparatus of claim 1 further comprising:
an ad server, coupled to comfort zone definition component, configured to place an ad certifying a user-generated page.
9. The computer system apparatus of claim 1 further comprising:
a moderation service component that provides an automated assessment of the content to determine if user input is required to classify the content and if the user input is required, flags the content for user input; and
the result aggregator accepts user input in the form of a user classification and uses the user classification to adjust the raw scores prior to aggregating the raw scores into an aggregate score.
10. The computer system apparatus of claim 9 further comprising:
a knowledge database configured to store universal resource locators (URLs), black and white lists, classifiers and a dictionary; and
a moderation service component further configured to remove content from the web page servers based on knowledge stored in the knowledge database.
11. A computer system for classifying digital content by using aggregate scoring, the computer system comprising:
a computer processor;
a memory, coupled to the computer processor, wherein the memory stores a program that classifies digital content by using aggregate scoring and when executed by the computer processor causes the processor to:
filter content received at web page servers of a multi-user website based on selected criteria to preliminarily identify content for specific analysis by a semantics parser function;
perform a specific type of analysis of supplied content and to output a content analysis result by a plurality of content analysis engines function;
receive content for specific analysis from the semantics parser function and couple to the plurality of content analysis engines using a grid load balancer function, with the grid load balancer function assigning content to different content analysis engines based on content identifications of the semantics parser; and
assign raw scores to content analysis results output from each content analysis engine, adjust aggregate scores based on user input, and aggregate the raw scores into an aggregate score for a particular grouping of content analysis results that reflects, at least in part, scores from two or more content analysis results by a result aggregator.
12. The computer system of claim 11 , further comprising causing the processor to:
provide an automated assessment of the content by a moderation service component to determine if user input is required to classify the content and if the user input is required, the moderation service component flags the content for user input; and
accept by the result aggregator user input in the form of a user classification and the result aggregator uses the user classification to adjust the raw scores prior to aggregating the raw scores into an aggregate score.