A novel modular connectionist architecture is presented in which the networks composing the architecture compete to learn the training patterns. As a result of the competition, different networks learn different training patterns and, thus, learn to compute different functions. The architecture performs task decomposition in the sense that it learns to partition a task into two or more functionally independent vii tasks and allocates distinct networks to learn each task. In addition, the architecture tends to allocate to each task the network whose topology is most appropriate to that task, and tends to allocate the same network to similar tasks and distinct networks to dissimilar tasks. Furthermore, it can be easily modified so as to...

We describe a distributed model of information processing and memory and apply it to the representation of general and specific information. The model consists of a large number of simple processing elements which send excitatory and inhibitory signals to each other via modifiable connections. Information processing is thought of as the process whereby patterns of activation are formed over the units in the model through their excitatory and inhibitory interactions. The memory trace of a processing event is the change or increment to the strengths of the interconnections that results from the processing event. The traces of separate events are superimposed on each other in the values of the connection strengths that result from the entire set of traces stored in the memory. The model is applied to a number of findings related to the question of whether we store abstract representations or an enumeration of specific experiences in memory. The model simulates the results of a number of important experiments which have been taken as evidence for the enumeration of specific experiences. At the same time, it shows how the functional equivalent of abstract representations—prototypes, logogens

Despite the fact that complex visual scenes contain multiple, overlapping objects, people perform object recognition with ease and accuracy. One operation that facilitates recognition is an early segmentation process in which features of objects are grouped and labeled according to which object they belong. Current computational systems that perform this operation are based on predefined grouping heuristics. We describe a system called MAGIC that learns how to group features based on a set of presegmented examples. In many cases, MAGIC discovers grouping heuristics similar to those previously proposed, but it also has the capability of finding nonintuitive structural regularities in images. Grouping is performed by a relaxation network that attempts to dynamically bind related features. Features transmit a complex-valued signal (amplitude and phase) to one another; binding can thus be represented by phase locking related features. MAGIC's training procedure is a generalizatio...

Neurons in the ventral stream of the primate visual system exhibit responses to the images of objects which are invariant with respect to natural transformations such as translation, size, and view. Anatomical and neurophysiological evidence suggests that this is achieved through a series of hierarchical processing areas. In an attempt to elucidate the manner in which such representations are established, we have constructed a model of cortical visual processing which seeks to parallel many features of this system, specifically the multi-stage hierarchy with its topologically constrained convergent connectivity. Each stage is constructed as a competitive network utilising a modified Hebb-like learning rule, called the trace rule, which incorporates previous as well as current neuronal activity. The trace rule enables neurons to learn about whatever is invariant over short time periods (e.g. 0.5 s) in the representation of objects as the objects transform in the real world. The trace ru...

A bst ract and-egg problem can be solved by using a cooperative computation in One way to achieve viewpoint-invariant shape recognition is to impose a canonical, object-based frame of reference on a shape and to describe the positions, sizes and orientations of the shape's features relative to the imposed frame. This compulation can be implemented in a parallel network of neuron-like processors, but the network has a tendency to make errors of a peculiar kind: When presented with several shapes it sometimes perceives one shape in tlie position of another. The parameters can be carefully tuned to avoid these "illusory conjunctions " in normal circumstances, but they reappear if the visual input is replaced by a random mask before the network has settled down. Treisman and Schmidt (1982) have shown that people make similar errors. 1.

Introduction Studies of population coding, which explore how the activity of ensembles of neurons represent the external world, normally focus on the accuracy and reliability with which sensory information is represented. However, the encoding strategies used by neural circuits have undoubtedly been shaped by the way the encoded information is used. The point of encoding sensory information is, after all, to generate and guide behavior. The ease and efficiency with which sensory information can be processed to generate motor responses must be an important factor in determining the nature of a neuronal population code. In other words, to understand how populations of neurons encode we cannot overlook how they compute. Gain modulation, which is seen in many cortical areas, is a change in the response amplitude of a neuron that is not accompanied by a modification of response selectivity. Just as population coding is a ubiquitous form of information representation, gain modulati

Summary The development of the issue of binding as fundamental to neural dynamics has made possible recent advances in the modeling of di cult problems of perception and brain function. Among them is perceptual segmentation, invariant pattern recognition and one-shot learning. Also, longer-term conceptual developments that have led to this success are reviewed.

The basis function theory of spatial representations explains how neurons i n the parietal cortex can perform nonlinear transformations from sensory to motor coordinates. The authors present computer simulations showing that unilateral parietal lesions leading to a neuronal gradient in basis function maps can account for the behavior of patients with hemineglect, including (a) neglect in line cancellation and line bisection experiments; (b) neglect in multiple frames of reference simultaneously; (c) relative neglect, a form of what is sometime called object-centered neglect; and (d) neglect without optic ataxia. Contralateral neglect arises in the model because the lesion produces an imbalance in the salience of stimuli that is modulated by the orientation of the body in space. These results strongly support the basis function theory for spatial representations in humans and provide a computational model of hemineglect at the single-cell level. A unilateral lesion of the parieto-occipital cortex in humans often produces hemineglect (Heilman, Watson, & Valenstein, 1985; Pouget & Driver, 1999; Vallar, 1998), a neurologic syndrome characterized by a conspicuous inability to react or respond to stimuli presented in the hemispace contralateral to the lesion. For example, when asked to

We describe a general framework for modeling transformations in the image plane using a stochastic generative model. Algorithms that resemble the well-known Kalman filter are derived from the MDL principle for estimating both the generative weights and the current transformation state. The generative model is assumed to be implemented in cortical feedback pathways while the feedforward pathways implement an approximate inverse model to facilitate the estimation of current state. Using the above framework, we derive models for invariant recognition, motion estimation, and stereopsis, and present preliminary simulation results demonstrating recognition of objects in the presence of translations, rotations and scale changes. 1

We describe two models of attention that utilize probabilistic principles to compute task-relevant variables. In the first model, objects and visual scenes are represented iconically using spatial filters at multiple scales. A maximum likelihood-based approach is used to compute the location of a target in a given scene. The eye movements generated by such a strategy are shown to be similar to human eye movement patterns elicited during visual search in naturalistic scenes. The second model is based on the statistical concept of predictive coding. It assumes that top-down feedback from higher cortical areas conveys predictions of expected activity at lower levels while the errors in prediction are conveyed through feedforward connections. The model explains how multiple objects in a scene can be recognized sequentially without an explicit spotlight of attention. An extension of the model provides an interpretation of object-based versus spatial attention in terms of interactions between “what ” and “where ” networks in the visual pathway.