Posts Tagged ‘Artificial Intelligence’

Image credit: Inserm / P. Latron

This is possible through the development of a “simplified artificial brain” that reproduces certain types of so-called “recurrent” connections observed in the human brain. The artificial brain system enables the robot to learn, and subsequently understand, new sentences containing a new grammatical structure. It can link two sentences together and even predict how a sentence will end before it is uttered. More info: http://bit.ly/15soo00

Journal article: Real-Time Parallel Processing of Grammatical Structure in the Fronto-Striatal System: A Recurrent Network Simulation Study Using Reservoir Computing. PLoS ONE, 2013 http://bit.ly/Xj3Zqz


Read Full Post »

The company’s name isn’t Cyberdyne Systems and the system they’ve developed isn’t called SkyNet, but I still found the headline, “US Army Gets Artificial Intelligence” a little bit disturbing. Ariston Consulting and ai-one, have apparently got together to supply Uncle Sam’s foot soldiers with, “machine learning tools to test and evaluate military Command, Control, Communications, Computers, Cyber, and Intelligence Test Bed systems.”

The companies in questions have developed tools to enable developers of military code to build machine learning into their applications in order that they can recognize patterns and associations like a human does. It gets better. ai-one reckon that their software is better because it learns without any human intervention. They reckon that their AI software, “detects the inherent structure of data in very complex environments so computers can recognize patterns and associations.” That’s great, does it also send a robot army to destroy mankind?

Amongst the companies plans for the future are AI analysis of threat patterns in cyber security, cataloging radio signals and network monitoring and management. The companies make a lot of noise about how much this will save the tax-payer (10% according to the press release) but it doesn’t mention what happens when the AI invades global cyber systems, gains control of nuclear launch codes and begins Judgement Day. You heard it here first folks. I’m Reese. Sergeant Tech-Com, DN38416. Assigned to protect you. You’ve been targeted for termination.

Read Full Post »

Taking strides toward understanding how the brain processes stimuli to recognise images, researchers have figured out how to project neural activity on to a TV screen.

How do they do it?
UC Berkeley professor Jack Gallant and his team use MRI to track blood-flow changes in a subject’s primary visual cortex – the brain’s largest visual processing centre – as he or she watches a movie. The researchers then create a model of the visual cortex that matches the blood-flow pattern with the images the subject is viewing.

Algorithms are applied to compare the brain signals with a catalogue of about 5.000 hours of YouTube video. The images that most accurately correspond to the brain activity are compiled into a composite video that resembles the YouTube footage.

In this video, Gallant explains how they succeeded in decoding and reconstructing people’s dynamic visual experiences…

Read Full Post »