Deric Bownds (here) has a post on ideas of Ullman et al in a recent PNAS paper From simple innate biases to complex visual concepts. Here is the paper’s 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.
This research seems to illuminate the problem with new born learning – how much is knowledge of the world innate and how much is learned. For example, do we need an inherited language module to learn language? In the case of hands and gaze, the babies seem to need only a very simple concept/motivation to begin their learning – the concept of a ‘mover’ and the motivation to follow movers. A very small input of innate knowledge can start of baby off in learning if it is the right little bit.