Slashdot Mirror
Neuroscientists At MIT Developing DNI
coolphysco1010 wrote to discuss the possible development of a direct neural interface, ala 'The Matrix', that could eventually allow for instant object recognition. From the article: "Now, neuroscientists in the McGovern Institute at MIT have been able to decipher a part of the code involved in recognizing visual objects. Practically speaking, computer algorithms used in artificial vision systems might benefit from mimicking these newly uncovered codes ... In a fraction of a second, visual input about an object runs from the retina through increasingly higher levels of the visual stream, continuously reformatting the information until it reaches the highest purely visual level, the inferotemporal (IT) cortex. The IT cortex identifies and categorizes the object and sends that information to other brain regions."
I don't want to see the results when they start trying to recreate those nueral patterns in the monkeys brains. Honesty, to say that observing these kinds of patterns brings us any closer to injecting images directly into the brain, when we have so little technology to do that (knives and chemicals basically) is ludicrous. I suppose the writer, rather than the scientists, can probably take all the credit for that exaggeration.
Those are my principles. If you don't like them I have others. -Groucho Marx
The implications for using this technology to cure blindness (one day, obviously not immediately) are wonderful! This is the kind of thing science was really meant for - helping humans live better lives. Kudos to MIT!
OK I just read the Science article. What's interesting about it is that they got recordings from a large population of neurons in IT during object recognition and have some cool analyses of the kinds of information that can be extracted from the capture, e.g. how large a population of neurons you need to accurately identify the object, how well the neurons discriminated among the categories and generalized across the same image at different sizes and positions, etc.
Important to remember that these monkeys were trained on a limited stimulus set, so its not that you can tell what the monkey is looking at by loooking at the recordings without knowing it is one of these pre-trained items.
Yeah, cool, well, actually getting any sort of input to the brain seems to be a big part of, if not THE actual problem.
Jeez.
No, seriously, I just come here for the articles.
I'm getting a bit tired of MIT getting press for research that has already been done years ago. In this case in particular, see the Dobelle Institute: here , here , and here , for instance.
Seriously. Don't exacerbate the inflated delusions of these guys by pretending that their research is unique or "cutting-edge". Expect more of them.