DARPA & Raytheon Create AI Wiki to Database Events and Peoplehttp://www.activistpost.com/2012/07/darpa-raytheon-create-ai-wiki-to.html
In recent weeks, more and more articles have been appearing in mainstream sources announcing the arrival of compartmentalized forms of artificial intelligence. On June 26, I wrote an article titled, “New Intelligent Biometric Security Program Can Adapt To Human Behavior,” which dealt with the announcement by Biometric Technologies Laboratory that researchers have developed a biometric security program that is able to adapt to changing circumstances and make intelligent decisions regarding the information it receives.
Continuing with this trend, a more recent announcement has also been issued from Raytheon, the notorious American defense contractor, and the shadowy DARPA (Defense Advanced Research Project Agency) regarding the development of an artificial intelligence (AI) program designed to comb news sites and TV broadcasts and boil down the information contained therein to a single concise article for the intelligence agent launching the query.
Obviously, such technology is not being provided for the benefit or convenience of the general public. In fact, it is not being released to the public at all. The users of the new product will be none other than the Pentagon itself.
At first glance, the end results of the new systems resemble nothing more than brief Wikipedia articles. However, unlike Wikipedia, the articles are generated in almost real time in response to the queries made by the user regarding individuals, events, or organizations and the connections between them.
The articles thus consist of profiles of these organizations, events, and people, gleaned from a stunning variety of online and broadcast sources which are then provided in short concise summaries to the intelligence analyst who made the initial search.
In order to do this, however, the computers that generate the results must not only be able to gather information from a wide variety of English websites and broadcasts, but also an equally impressive amount of news sites from around the world and in other languages. Going further, the computers must be able to analyze, collate, and reproduce these results in a coherent format.
Of course, the latter half of the technology’s responsibility will require more than sophisticated computational abilities. Indeed, the ability to analyze information – sifting through colloquial voice, errors, and sarcasm – will require a type of intelligence which is unique to humans, i.e. the ability to understand what is being said as an interaction, complete with understanding of intention, as opposed to mere informational communication.
Nevertheless, as impossible as this may sound at first, the folks at DARPA and Raytheon seem to have managed to create just this type of system, along with the requisite capabilities.
The creators of the program have provided limited access to the technology via demonstrations and interviews such as the one provided to Technology Review. In the article entitled, “An Online Encyclopedia That Writes Itself,” David Talbot describes how the program works.
Talbot writes, “It starts by detecting an ‘entity’ – a name or organization, such as Boko Haram, accounting for a variety of spellings. Then it identifies other entities (events and people) that are connected to it, along with statements made by and about the subject.”
Talbot also quotes Sean Colbath, a senior scientist at BBN Technologies, who conducted the demonstration as stating, “It’s automatically extracting relationships between entities. . . . . . Here the machine has learned, by being given examples, how to put these relationships together and fill in those slots for you.”
Colbath continues by saying, “I could go and read 200 articles to learn about Bashar Al-Assad (the Syrian dictator). But I’d like to have a machine tell me about it.”
According to Talbot, DARPA’s program manager of the project, Bonnie Dorr, added that “the technology incorporates recent improvements in machine reading, enabling to do a better job of understanding when the same underlying event is described in multiple ways – such as ‘Joe is married to Sue’ and ‘Sue is Joe’s spouse’ – and to determine the sentiment implied in phrases like ‘really awesome.’”
Interestingly enough, the BBN project known as The BBN Broadcast Monitoring System, is actually being directed by Raytheon, as BBN is a wholly owned subsidiary of the defense contractor. Indeed, the new name of the company is now Raytheon BBN Technologies, if there were ever any doubt.
Yet even Raytheon is not the true source of the project. It is merely a contractor of the U.S. government by way of DARPA. The umbrella project of the new Surveillance Wiki is called the Machine Reading Program which itself is contained under the Machine Research Program.
According to the DARPA website states that the Machine Reading Program
aims to address this issue [“archaic” AI first order logic] by replacing expert and associated knowledge engineers with un-supervised or self-supervised learning systems that can ”read” natural text and insert it into AI knowledge bases (i.e., data stores especially encoded to support subsequent machine reasoning). If successful, the Machine Reading program will produce language-understanding technology that will automatically process text in timelines consistent with operational tempo.
With these recent announcements, it is clear that those in the upper reaches of Governments and major Corporations are rolling out Artificial Intelligence programs with increasing vigor. Obviously, the timing is right from the point of view of the technocratic elite to introduce such systems to the general public.
By acclimating citizens to such programs as the Machine Reading or BBN Broadcast Monitoring System, the stage then becomes set for even bigger programs that are no doubt already developed and, to some degree, already being implemented.
Although the systems described in David Talbot’s Technology Review article are presented as being in their infancy stages, we cannot be so naïve as to think that either DARPA or Raytheon would announce their new products or national security technology before they have been completed. Indeed, if history is anything to go by, then what has truly been developed is vastly more sophisticated than an AI Wiki.