GateFilter is based on AI (Artificial
Intelligence) Content Filtering technology, which scans the content of the site with
intelligent algorithms allowing the software to "decide" whether the subject
matter is appropriate prior to displaying it. This proprietary process developed by
iCognito Technologies Ltd. boasts an aggressive 99.2% accuracy rate, without
under-filtering or over-filtering.
back to top
Two filtering technologies commonly
exist, database filtering and keyword filtering.
Database filtering compares a requested URL (Uniform Resource Locator, i.e. a website) to
a list of categorized sites, blocking those contained in the database. The largest known
database has just over two million URLs, compiled by a committee of individuals who
examine websites to determine if they are offensive. However, the World Wide Web has
billions of web pages, and ten million new pages are added daily. Therefore, database
filtering commonly under filters content.
Keyword filtering reads the URL of the requested site and searches for certain words or
sequences of letters. This solution has received criticism for inaccuracy and over
filtering because the content surrounding the identified keyword(s) is not considered.
GateFilter differs from products on the market today because it does not rely upon
keywords or databases as its primary means of filtering. Instead, the primary filtering
technology is based upon AI (Artificial Intelligence) which previews the content of a
website, determines if it is inappropriate based on the engines experience and collected
data, and blocks the site if necessary.
back to top
Through central installation on the
WinGate server, GateFilter is seamlessly integrated into WinGate's GateKeeper.
Administration is facilitated through a simple tabbed interface where network-wide
settings are manipulated. The settings of GateFilter are protected through the
Administrative Login required to access GateKeeper.
back to top
Websites are categorized real-time
within GateFilter into two categories, sex and gambling. Categories are not based on an
existing database. However, once the site is categorized within the filtering engine, the
site category is cached in the memory to improve performance. Each category is accompanied
by settings specific to that category, allowing the level of filtering to vary dependant
on the category. Additional GateFilter categories will be available in the near future.
back to top
GateFilter permits flexibility in the
level of filtering through user-defined settings. The Administration may set the filtering
specifications from 0% to 100%, customizing the amount of filtering needed. The percentage
selected by category correlates to the certainty that the filtering engine places on its
decision to block the site. For example, if the low percentage of 10% is chosen, the
engine will block any website that it is 10% certain of containing objectionable material.
If, however, a higher percentage of 90% is chosen, the engine will only block websites
that it is nearly positive of containing objectionable material. A lower percentage
results in more rigid filtering. Whereas, a greater percentage will require additional
time for the filtering engine to study each website.
back
to top
Known websites containing
objectionable content may be added to a static URL list within the filtering engine. If
the website is requested, GateFilter will automatically block the site. Specific URL
filtering provides a secondary layer of protection beyond the AI (Artificial Intelligence)
filter.
back to top
As a plug-in to WinGate, GateFilter
provides content filtering for the entire network. GateFilter eliminates the requirement
for multiple installations (and licenses) of workstation software, while minimizing the
security threat posed if individual client computer filtering is disabled.
back to top |
Current verion of GateFilter is 1.0