Object Recognition and Sensitive Periods: A Computational Analysis of Visual Imprinting

Abstract

Using neural and behavioral constraints from a relatively simple biological visual system, we evaluate the mechanism and behavioral implications of a model of invariant object recognition. Evidence from a variety of methods suggests that a localized portion of the domestic chick brain, the intermediate and medial hyperstriatum ventrale (IMHV), is critical for object recognition. We have developed a neural network model of translation-invariant object recognition that incorporates features of the neural circuitry of IMHV, and exhibits behavior qualitatively similar to a range of findings in the filial imprinting paradigm. We derive several counter-intuitive behavioral predictions that depend critically upon the biologically derived features of the model. In particular, we propose that the recurrent excitatory and lateral inhibitory circuitry in the model, and observed in IMHV, produces hysteresis on the activation state of the units in the model and the principal excitatory neurons in IMHV. Hysteresis, when combined with a simple Hebbian covariance learning mechanism, has been shown in this and earlier work (Foldidk 1991; O'Reilly and McClelland 1992) to produce translation-invariant visual representations. The hysteresis and learning rule are responsible for a sensitive period phenomenon in the network, and for a series of novel temporal blending phenomena. These effects are empirically testable. Further, physiological and anatomical features of mammalian visual cortex support a hysteresisbased mechanism, arguing for the generality of the algorithm.

Predictive Coding and the Slowness Principle: An Information-Theoretic Approach

“…Various paradigms for invariance learning have been proposed that exploit this observation (Földiak, 1991;Wallis & Rolls, 1997;O'Reilly & Johnson, 1994;Stone & Bray, 1995;Einhäuser, Hipp, Eggert, Körner, & König, 2005). As these paradigms extract the slowly varying components of sensory signals, we will refer to this approach as the slowness principle (Wiskott & Sejnowski, 2002), in related literature also called temporal coherence or temporal stability principle (Einhäuser et al, 2005;Hurri & Hyvärinen, 2003;Wyss, König, & Verschure, 2006).…”

Slow Feature Analysis: A Theoretical Analysis of Optimal Free Responses

“…This approach to unsupervised learning of invariant representations has been taken by a number of researchers since the early 1990s (Földiák, 1991;Mitchison, 1991;Becker & Hinton, 1992;O'Reilly & Johnson, 1994;Stone & Bray, 1995;Wallis & Rolls, 1997;Peng, Sha, Gan, & Wei, 1998;Kayser, Einhäuser, Dümmer, & König, 2001;Wiskott & Sejnowski, 2002) and an earlier description of the principle can be found in Hinton (1989). Computational models based on the principle of temporal slowness have been quite successful in learning invariances in a number of contexts (see references above) and in reproducing receptive field properties of the primary visual cortex (Kayser et al, 2001;Berkes & Wiskott, 2002).…”

The Development of Animal Behavior: From Lorenz to Neural Nets

“…put forward by Bateson and Horn (1994). A similar model was proposed by O'Reilly and Johnson (1994). Figure 5 is a simple diagram of the basic scheme of Bateson and Horn's model.…”