A class of adaptive resonance theory (ART) models for learning, recognition, and prediction with arbitrarily distributed code representations is introduced. distributed ART neural networks combine the stable fast lear...
详细信息
A class of adaptive resonance theory (ART) models for learning, recognition, and prediction with arbitrarily distributed code representations is introduced. distributed ART neural networks combine the stable fast learning capabilities of winner-take-all ART systems with the noise tolerance and code compression capabilities of multilayer perceptrons. With a winner-take-all code, the unsupervised model dART reduces to fuzzy ART and the supervised model dARTMAP reduces to fuzzy ARTMAP. With a distributed code, these networks automatically apportion learned changes according to the degree of activation of each coding node, which permits fast as well as slow learning without catastrophic forgetting. distributed ART models replace the traditional neural network path weight with a dynamic weight equal to the rectified difference between coding node activation and an adaptive threshold Thresholds increase monotonically during learning according to a principle of atrophy due to disuse. However, monotonic change at the synaptic level manifests itself as bidirectional change at the dynamic level, where the result of adaptation resembles long-term potentiation (LTP) for single-pulse or low frequency test inputs but can resemble long-term depression (LTD) for higher frequency test inputs. This paradoxical behavior is traced to dual computational properties of phasic and tonic coding signal components. A parallel distributed match-reset-search process also helps stabilize memory. Without the match-reset-search system, dART becomes a type of distributed competitive learning network. (C) 1997 Elsevier Science Ltd.
The mammalian superior colliculus is involved in the transformation of sensory signals into orienting behaviors. Sensory and motor signals are integrated in the colliculus to produce movements of the eyes, head, and n...
详细信息
The mammalian superior colliculus is involved in the transformation of sensory signals into orienting behaviors. Sensory and motor signals are integrated in the colliculus to produce movements of the eyes, head, and neck. While there is a considerable amount of information available on the afferent and efferent connections of the colliculus, almost nothing is known about its intrinsic circuitry, particularly that of its deepest layers. It is likely that intrinsic connections in these deeper layers of the colliculus participate in the sensory-motor transformations leading to orienting movements. In this study, we used the neuroanatomical tracer biocytin to label small groups of neurons in the deeper layers of the cat superior colliculus and examine the distribution of their axons and terminals. We found a broadly distributed network of intrinsic projections throughout the deep layers of the superior colliculus. While the majority of terminals were found in a 1-2 mm radius around the injection site, labeled terminals were found throughout the deep layers of the colliculus up to 5 mm from the injection site. In addition, these injections sometimes labeled terminals in the superficial tectum. Extensive projections were demonstrated by the more superficial injections, but few terminals were found when injections were confined to the deepest layers of the colliculus. There was no evidence of anisotropy in the distribution of terminals from injections made at different rostrocaudal or mediolateral locations;neurons located in any one region in the colliculus could potentially influence any other region. This network of intrinsic connections in the cat superior colliculus could provide a means for deeper-layer efferent neurons to associate, and to modulate or coordinate their output. Interneurons could also provide a substrate for mutual inhibition between neurons at the rostral pole of the colliculus that are active during fixation, and more caudally located neurons whose
A new context-sensitive neural network, called an EXIN (excitatory + inhibitory) network, is described. EXIN networks self-organize in complex perceptual environments, in the presence of multiple superimposed patterns...
详细信息
A new context-sensitive neural network, called an EXIN (excitatory + inhibitory) network, is described. EXIN networks self-organize in complex perceptual environments, in the presence of multiple superimposed patterns, multiple scales, and uncertainty, The networks use a new inhibitory learning rule, in addition to an excitatory learning rule, to allow superposition of multiple simultaneous neural activations ( multiple winners), under strictly regulated circumstances, instead of forcing winner-take-all pattern classifications. The multiple activations represent uncertainty or multiplicity in perception and pattern recognition. Perceptual scission (breaking of linkages) between independent category groupings thus arises and allows effective global context-sensitive segmentation, constraint satisfaction and exclusive credit attribution. A Weber Law neuron growth rule lets the network learn and classify input patterns despite variations in their spatial scale, Applications of the new techniques include segmentation of superimposed auditory or biosonar signals, segmentation of visual regions, and representation of visual transparency.
This paper reviews evidence that the superior colliculus (SC) of the midbrain represents visual direction and certain aspects of saccadic eye movements in the distribution of activity across a population of cells. Acc...
详细信息
This paper reviews evidence that the superior colliculus (SC) of the midbrain represents visual direction and certain aspects of saccadic eye movements in the distribution of activity across a population of cells. Accurate and precise eye movements appear to be mediated, in part at least, by cells of the SC that have large sensory receptive fields and/or discharge in association with a range of saccades. This implies that visual points or saccade targets are represented by patches rather than points of activity in the SC. Perturbation of the pattern of collicular discharge by focal inactivation modifies saccade amplitude and direction in a way consistent with distributed coding. Several models have been advanced to explain how such a code might be implemented in the colliculus. Evidence related to these hypotheses is examined and continuing uncertainties are identified.
暂无评论