Computer scientists from Carnegie Mellon University are developing and using online games to train computers visual system according to this article from the Pittsburgh Post-Gazette.

An example game is Peekaboom, which is used online by teams of two players. The first one, designated as “Peek,” sees on his screen an image — initially empty — and a word that describes the image or one element of the image. The second one, named “Boom,” gradually reveals the image or gives hints to “Peek” until he correctly guesses the word associated to the image. A computer program can use this information to learn how to better categorize and identify objects visually. Essentially, humans are helping to train computer vision systems.

This reminds me of the method people use to break CAPTCHAs. A Completely Automated Public Turing Test to Tell Computers and Humans Apart is often used to protect web forms from automated use. To break a CAPTCHA, one simply needs to redirect the CAPTCHA image that the script comes across - to a human of course. The most common way to do this, is to ask a human to evaluate a CAPTCHA image before giving them access to free pornography. Clever, but extremely simply, and I would guess it is quite effective.

Why I blog this? I believe we will see many more ‘teaching systems’ where computer software ‘learns’ by working with humans. Why? Because humans already know how to learn. We can categorize incoming sensory information and manipulate it abstractly with symbols. And, we know instinctively how to teach learning when we raise children. I quite expect to one day raise a personal A.I. using human training and attention.

via FutureWire via Unmediated via SmartMobs

The human brain has access to a massive amount of visual sensory data. Researchers have conducted some interesting studies to figure out how our brain decides what small window of that data we direct our attention to. The researchers determined that when we notice an object appearing, we pay attention to it for an extended period of time (1/3 of a second). But, when we notice an object disappearing we immediately direct our attention elsewhere.

[The appearing object effect] was first observed by Michael Posner — if an object appears in one part of our field of view, it temporarily delays our ability to detect another object appearing near it. The effect begins about a third of a second after the first object appears and lasts about a second. If the second object appears sooner than that, we actually notice it quicker. Subsequent research revealed that the effect became progressively smaller at greater and greater distances from the spot where the first object appeared — surprisingly, we’re quicker to spot other objects appearing farther away from the original object.

See Cognitive Daily for the full article, including the research on disappearing objects.

Recent research at the California Institute of Technology has confirmed that an area of the human brain, the ventrolateral prefrontal cortex (vPF), is involved in the planning stages of movement. The planning stages of movement happen during the instantaneous flicker of time when we contemplate moving a limb. This has implications for the development of brain-machine interfaces for the paralyzed as well as for able-bodied people who may seek to augment themselves with such technology. According the press release, the work currently appears in the online version of Nature Neuroscience. I do not subscribe to Nature so unfortunatly I can not access this paper.

“We were looking for the brain regions that may be contributing to planned movements. And what I was able to show is that a part of the brain called the ventrolateral prefrontal cortex is indeed involved in planning these movements.” Just by analyzing the brain activity from the implanted electrodes using software algorithms that he wrote, Rizzuto was able to tell with very high accuracy where the target was located while it was on the screen, and also what direction the patient was going to reach to when the target wasn’t even there.

Why I blog this? Practical consumer/amateur level brain-computer interfaces, with at least rudimentary functionality, are within reach. Obviously implanted electrodes are not practical for most people, but research is advancing in all directions. So, if we do get external BCI’s, even if they act simply as a new form of mouse, they will still be very hot technology to play with. I try to keep somewhat up to date on developments in this field.

via Caltech Press Release - Scientists Discover What You Are Thinking, March 16, 2005
via KurzweilAI.net

How much can we understand about human motion using only simple cues? Take a look for yourself using BioMotionLab1.6, a flash-based visual demonstration developed by an international team of researchers headed by Prof. Dr. Nikolaus Troje. Adjust four different sliders (male/female, heavy/light, nervous/relaxed, and happy/sad) to change the motion of the dots. It is amazing how much the human mind can ‘read’ from the motion, even though only minimal data is presented. I played around with it for a few minutes and found their simulation to be a fairly elegant examination of human movement using such a small number of points.

Dr. Troje holds a Canadian Research Chair in Vision and Behavioural Sciences. The Biomotion lab operates out of two locations: Queen’s University in Kingston, Ontario and Ruhr-University in Bochum, Germany.

From the Biomotion lab site:

We are working on several aspects of visual perception and cognition. Our major interest is focused on questions concerning the biology and psychology of social recognition. That is: conspecific recognition, gender recognition, individual recognition, recognition of an agent’s actions, intentions, and emotions and personality traits.

In the past, we had mainly worked on the psychophysics and modelling of human face recognition. More recently, our focus shifted towards perception of biological motion as a major source of social information.

The goal of our current work is to provide a solid basis for the description, analysis and synthesis of animate motion patterns. We want to achieve a comprehensive understanding of the information transmitted through biological motion, its perception and underlying neuronal mechanisms. In addition to human psychophysics, we are using the pigeon as a model for ethological and neurophysiological investigations in the context of courtship behaviour, social learning, and social fascilitation.

The procedure they used to create the display will be described in detail in a forthcoming paper submitted to the Journal of Vision. Troje, N. F. (submitted) Decomposing biological motion: A framework for the analysis and synthesis of human gait patterns.

Why I blog this? I referenced it in my Cybernetics and Society (STV205) class, but was unable to actually demonstrate the visualization because I did not have a computer available at the time. To anyone from class who was interested - Enjoy :)