Skynet Awaits: IBM Tests AI That Replicates Human Brain Activity
The human brain is this wonderfully complex organic mass of nerve fibers and electrical impulses, and as much as we know about it today from an anatomical standpoint, there are many mysteries that we have yet to solve. It's what we don't know that makes true artificial intelligence so difficult, though it's also alluring, which is one reason why IBM is getting involved with machine learning.
Palm founder Jeff Hawkins has invested his own time and money into mapping out the human brain in hopes of advancing AI technology through his company Numenta. You've probably never heard of Numenta, which speaks to what little impact it's had on the industry, but it's getting a partner in IBM, which has formed a research group to work on Numenta's learning algorithms at its Almaden research lab in San Jose, California, MIT Technology Review reports.
Why Numenta? IBM researcher and project leader Winfried Wilcke has gone on record saying that Numenta's software is closer to being a biological reality than any other machine learning software out there -- high praise, considering companies like Google also use machine learning software. One thing Wilcke likes about Numenta's algorithm is that it can get right to work on a problem rather than having to be trained with example data like most other machine learning software.
At present, those algorithms are being tested for tasks like interpreting satellite imagery of crops and to detect early warning signs of mechanical failures in data from pumps and other machinery. Where it goes from there depends partially on what IBM can bring to the table and what advances are made. As it exists, the algorithms are designed to recreate the behavior of repeating circuits of around 100 neurons found in the outer layer of the brain called the neocortex.
"Our goal is not to be biologically inspired; I want to re-create exactly," says Hawkins. “This is how you would really build a machine intelligence."
Palm founder Jeff Hawkins has invested his own time and money into mapping out the human brain in hopes of advancing AI technology through his company Numenta. You've probably never heard of Numenta, which speaks to what little impact it's had on the industry, but it's getting a partner in IBM, which has formed a research group to work on Numenta's learning algorithms at its Almaden research lab in San Jose, California, MIT Technology Review reports.
Image Source: Flickr (H. Michael Karshis)
Why Numenta? IBM researcher and project leader Winfried Wilcke has gone on record saying that Numenta's software is closer to being a biological reality than any other machine learning software out there -- high praise, considering companies like Google also use machine learning software. One thing Wilcke likes about Numenta's algorithm is that it can get right to work on a problem rather than having to be trained with example data like most other machine learning software.
At present, those algorithms are being tested for tasks like interpreting satellite imagery of crops and to detect early warning signs of mechanical failures in data from pumps and other machinery. Where it goes from there depends partially on what IBM can bring to the table and what advances are made. As it exists, the algorithms are designed to recreate the behavior of repeating circuits of around 100 neurons found in the outer layer of the brain called the neocortex.
"Our goal is not to be biologically inspired; I want to re-create exactly," says Hawkins. “This is how you would really build a machine intelligence."
