Psychology Magazine

How a Simple Innate Bias Might Guide Visual Learning

By Deric Bownds

For years computer programmers have been trying to design algorithms that even remotely approach the ability of young infants in their first few months of life to rapidly learn to recognize complex objects and events in the their visual input, particularly events like hand movements and gaze direction. Even the most powerful probabilistic learning models, as well as connectionist and dynamical models, do not result by themselves in automatically learning about hands, detecting them, paying attention to what they are doing, and using them to make inferences and predictions. Ullman et al. develop a model that incorporates a plausible innate or early acquired bias, based on cognitive and perceptual findings, to detect “mover” events. It leads to the automatic acquisition of increasingly complex concepts and capabilities, which do not emerge without domain-specific biases. After exposure to video sequences containing people performing everyday actions, and without supervision, the model develops the capacity to locate hands in complex configurations by their appearance and by surrounding context and to detect direction of gaze. Here is their abstract:

Early in development, infants learn to solve visual problems that are highly challenging for current computational methods. We present a model that deals with two fundamental problems in which the gap between computational difficulty and infant learning is particularly striking: learning to recognize hands and learning to recognize gaze direction. The model is shown a stream of natural videos and learns without any supervision to detect human hands by appearance and by context, as well as direction of gaze, in complex natural scenes. The algorithm is guided by an empirically motivated innate mechanism—the detection of “mover” events in dynamic images, which are the events of a moving image region causing a stationary region to move or change after contact. Mover events provide an internal teaching signal, which is shown to be more effective than alternative cues and sufficient for the efficient acquisition of hand and gaze representations. The implications go beyond the specific tasks, by showing how domain-specific “proto concepts” can guide the system to acquire meaningful concepts, which are significant to the observer but statistically inconspicuous in the sensory input.

You Might Also Like :

Back to Featured Articles on Logo Paperblog

These articles might interest you :

  • How the Brain Shifts Between Purpose and Habit.

    Gremel et al. do a nice explication of the brain's upstairs/downstairs story, illustrating with direct electrical recordings how the mouse brain switches betwee... Read more

    The 21 August 2013 by   Deric Bownds
  • How We Work: 'brain Waves' Versus Modern Phrenology

    Alexander et al. have analyzed data from magnetoencephalogram (MEG), electroencephalogram (EEG) and electrocorticogram (ECoG), focusing on globally synchronous... Read more

    The 29 May 2013 by   Deric Bownds
  • How We Listen to Music...

    Adam Gopnik has a very nice essay in The New Yorker on the mysteries of sound and the quest for 3-D recording. I was struck by his description of how the way... Read more

    The 14 February 2013 by   Deric Bownds
  • How Mindfulness Meditation Works in the Brain - a Model

    Kerr et al make some interesting speculations. Their article contains some useful summary graphics. Using a common set of mindfulness exercises, mindfulness... Read more

    The 20 February 2013 by   Deric Bownds
  • How Procrastination Gets Things Done...

    John Tierney writes a delightful piece on not getting some things done. He notes several authors' comments on procrastination: The key to productivity, John... Read more

    The 22 January 2013 by   Deric Bownds
  • How I Quit Gambling

    Projectile vomiting can be your friend. I never should have found myself inside casinos in the first place. As a former alcoholic, cigarette smoker, and drug... Read more

    The 08 February 2013 by   Dirkh
  • How Environments Talk to Genes.

    The January issue of Nature Neuroscience has some fascinating articles on gene-environment interactions. Vassoler et al. report that in rats paternal cocaine us... Read more

    The 18 January 2013 by   Deric Bownds

Add a comment