February 9th, 2008
Search Technology Is A Precursor To Artificial Intelligence
posted by Midwinter
Many of you who read our content reguarly, especially those of you who subscribed to Surrch.eu; know that Grant Midwinter has their own search technology. Our search crawlers and engines are highly adapatable and a lot of our work focuses on taking search functionality and adapating it for a useful purpose other than the straight Google or Yahoo style user search engine.
Google has the right idea, I believe, when they profess a goal of indexing all the world’s information. However, they seemed to have missed the part where they utilise the world’s information for a higher purpose. As the old adage goes, “it’s not what you have, it’s what you do with it that counts”. The main use of having large quantities of automatically gathered data is of course for pattern analysis. This may sound pretty boring until you realise that the kind of patterns we’re talking about are those that the human brain is able to spot so easily and we take for granted in the course of our daily lives. An interesting pattern that the human brain is able to observe (some better than others) is the social factors that make someone popular, or well liked. This is why search crawling, and analysis, is so important a step to generating a true Artificial Intelligence.
I theorise, that anything associated with humanity, is actually down to the advanced pattern recognition available within our brains. Even learning language is a repeated exposure to stimuli until we understand the associative context of a word such as ‘chair’ to its inanimate object - or for a more difficult scenario; the word love to the associated chemical batter our brains and bodies are flooded with in fine balance.
It is to this end that Grant Midwinter is now in the business of creating more advanced applications based on search and tending towards AI. Our first publically available sample will be an anti-piracy tool unlike any existing application on the market. Let me give you a bit of background here:
With the new trend of uploading copyrighted videos of tv shows and movies to open websites, such as Stage6.com and Veoh.com, a new breed of internet forum has emerged. These forums have lurkers just like any other but they also have ‘link hunters’. The link hunters find links to the copyrighted videos and post them on the forum for others to view. So technically they are breaking no laws, and the onus is on the video hosting site to remove the infringing content.
Link hunters have the advantage though, they have an army of human brains who are good at spotting patterns, knowing what is and isn’t popular content. They’re also pretty darn good at knowing what is infringing on copyright and how to post links to obfuscate their true locations from simple machines. So we’re going to use our semantic inference engine and crawler technology, adapt it, so that it will allow you to list all the links currently online of infringing content on the major video sites. The MPAA, RIAA and other such entities can then use it to post their DMCA (Digital Millenium Copyright Act) takedown notices and have the infringing content removed. Of course they’ll have to automate the process if they want to take down the material more quickly than it can be put up by the legions of link posters out there. It’s unlikely we’ll provide an API for this because it’s a test subject for us of the capability of our unique technology.
If you’re interested in how our engines work, or any of our thoughts on search and AI then as always just ask and we’ll do our best to give you answers. Look for the Link Hunter Articifical Intelligence on release day [UPDATE: Release delayed, more soon]