InBoxer, a new anti-spam filter based on Bayesian statistics and other state-of-the art techniques, debuted for public testing today. Unlike traditional spam fighters, InBoxer does not depend on rules or lists that can be regularly beaten by spammers. Instead, it uses harder to beat language technologies designed to remove more than 99 percent of spam and unwanted junk email.BOXBOROUGH, Mass. (PRWEB) June 25, 2003 InBoxer, a new anti-spam filter based on Bayesian statistics and other state-of-the art techniques, debuted for public testing today. Unlike traditional spam fighters, InBoxer does not depend on rules or lists that can be regularly beaten by spammers. Instead, it uses harder to beat language technologies designed to remove more than 99 percent of spam and unwanted junk email.InBoxer is designed by Audiotrieve, a company founded by industry veterans from Dragon Systems, the one-time speech recognition and language technology industry leader. The new InBoxer software is based on an analysis of the language used in email messages. It learns by example from the users mail folders and creates a unique personal filter for "good" and "junk" emails based on each users personal interests. Because each users filter is customized, spammers cannot design one message that will get through to everyone.Anti-spam software is getting better. However, no matter how good traditional spam stopping solutions become, these filters are doomed to failure because the spammers can outsmart them, said Roger Matus, Chief Executive of Audiotrieve. Most filters are based on rules and lists sent from the software vendor to its customers. Unfortunately, the spammers get the new rules at the same time and quickly bypass them. The result is an ongoing race between the traditional vendors and the spammers that the vendors cannot win.InBoxer compares each new email message to the personal models for good and junk email that it built. It is not enough for spammers to bypass the junk filter, each message must also be allowed by the good filter. Questionable messages are placed in a review folder and messages are never automatically deleted. Spammers also never receive centralized updates because each users filter is updated when new mail arrives. InBoxer does not require an update subscription to remain accurate.U.S. Supreme Court Justice Potter Stewart once said that he could not define pornography, but he said that I know pornography when I see it. It is possible that the same could be said about defining spam. It is hard to create a set of rules that would be accepted by everyone to define spam, but each individual may know what he or she considers spam when he or she sees it. That means the best way to categorize spam is for each person to categorize it by example, said Matus. Not all Bayesian filters are the same. Significant expertise is needed to build an accurate system. InBoxer is from a company founded by veterans of the speech and language industry who understand the importance of this process. InBoxer is designed for the millions of Microsoft Outlook (2000, 2002, XP) users under Windows. Users may try the beta test software for 21 days for free at www.inboxer.com. About Audiotrieve, LLC and InBoxerAudiotrieve, LLC (www.audiotrieve.com) creates products and services that sort, index and retrieve information that cannot be found easily using traditional means. InBoxer (www.inboxer.com), separates the email users want from the email they dont. Not just another spam fighter, it utilizes Bayesian statistics (www.inboxer.com/2faqbayesian.html) and language industry techniques to protect an Inbox. Planned offerings include services that help search engines, such as Yahoo! or Google, find the right content to a query from within an audio or video stream.
