This paper describes NETtalk, a class of massively-parallel network systems that learn to convert English text to speech. The memory representations for pronunciations are learned by practice and are shared among many processing units. The performance of NETtalk has some similarities with observed human performance. (i) The learning follows a power law. (;i) The more words the network learns, the better it is at generalizing and correctly pronouncing new words, (iii) The performance of the network degrades very slowly as connections in the network are damaged: no single link or processing unit is essential. (iv) Relearning after damage is much faster than learning during the original training. (v) Distributed or spaced practice is more effective for long-term retention than massed practice. Network models can be constructed that have the same performance and learning characteristics on a particular task, but differ completely at the levels of synaptic strengths and single-unit responses. However, hierarchical clustering techniques applied to NETtalk reveal that these different networks have similar internal representations of letter-to-sound correspondences within groups of processing units. This suggests that invariant internal representations may be found in assemblies of neurons intermediate in size between highly localized and completely distributed representations.

Present neural models of classical conditioning all suffer from the same shortcoming: local representation of information (therefore, very precise neural prewiring is necessary). As an alternative we develop two neural models of classical conditioning which rely on distributed representations of information. Both models are of the Hopfield type. In the first model the existence of transmission delays is used to store temporal relations. The second model is based on interactions between spatially separated neural fields. Using tools from statistical mechanics we show that behavioural constraints can be met only if the Hebb rule is extended with inter- or intrasynaptic competition. 2 3 1. Introduction Connectionism has redirected the attention of cognitive scientists to learning and to the neural substrate in which cognitive processes are implemented. Conditioning has become an important field in which ideas from neural networks, behavioural science and neurophysiology are combined. ...

Abstract not found

This paper presents a new approach to neural modeling based on the idea of using an automated method to optimize the parameters of a synaptic learning rule. The synaptic modification rule is considered as a parametric function. This function has local inputs and is the same in many neurons. We can use standard optimization methods to select appropriate parameters for a given type of task. We also present a theoretical analysis permitting to study the generalization property of such parametric learning rules. By generalization, we mean the possibility for the learning rule to learn to solve new tasks. Experiments were performed on three types of problems: a biologically inspired circuit (for conditioning in Aplysia), Boolean functions (linearly separable as well as non linearly separable) and classification tasks. The neural network architecture as well as the form and initial parameter values of the synaptic learning function can be designed using a priori knowledge.

Biological mechanisms are adapted to the exigencies of the functions they serve. The function of memory is to carry information forward in time. The function of learning is to extract from experience properties of the environment likely to be useful in the determination of future behavior. These different functions lead to different manifestations of the biological principle of adaptive specialization. Adaptive Specialization in the Memory Mechanism We do not know what the neurobiological mechanism of memory is, so we cannot say how it is adapted to its function. We can, however, look at other mechanisms that serve the same function to see what adaptations they suggest we may find in a mechanism with this function. One such mechanism is computer memory; another is DNA, the molecular mechanism for carrying hereditary information from one generation to the next. Both of these “memory ” mechanisms suggest two exigencies and two principles: The exigencies are thermodynamic stability and high density. The principles are that information is information and that information conveyance requires a code. Thermodynamic Stability

Abstract not found

Introduction Reinforcement learning (RL) studies the ways animals learn about the events in their environment that lead to rewards and punishments, and use this information to choose their actions appropriately. RL is minimally supervised since animals are not told explicitly what actions to perform in various situations, but rather must work this out for themselves on the basis of evaluative information. RL is one of the cases in which it is possible to derive computational, psychological and neurobiological constraints, and in which there are theories that satisfy many of the constraints at all the levels. Various neural systems have been studied from the perspective of RL. One is the cerebellum, a structure implicated in eyeblink conditioning, in which animals learn to shut their eyes just in advance of disturbances such as airpuffs or weak shocks that are signalled by cues. Another is the primate midbrain dopaminergic system and its targets. We focus on the latter, togeth

We draw together and develop previous timing models for a broad range of conditioning phenomena to reveal their common conceptual foundations: First, conditioning depends on the learning of the temporal intervals between events and the reciprocals of these intervals, the rates of event occurrence. Second, remembered intervals and rates translate into observed behavior through decision processes whose structure is adapted to noise in the decision variables. The noise and the uncertainties consequent upon it have both subjective and objective origins. A third feature of these models is their time-scale invariance, which we argue is a deeply important property evident in the available experimental data. This conceptual framework is similar to the psychophysical conceptual framework in which contemporary models of sensory processing are rooted. We contrast it with the associative conceptual framework.

Biological mechanisms are adapted to the exigencies of the functions they serve. The function of memory is to carry information forward in time. The function of learning is to extract from experience properties of the environment likely to be useful in the determination of future behavior. These different functions lead to different manifestations of the biological principle of adaptive specialization. Adaptive Specialization in the Memory Mechanism We do not know what the neurobiological mechanism of memory is, so we cannot say how it is adapted to its function. We can, however, look at other mechanisms that serve the same function to see what adaptations they suggest we may find in a mechanism with this function. One such mechanism is computer memory; another is DNA, the molecular mechanism for carrying hereditary information from one generation to the next. Both of these “memory ” mechanisms suggest two exigencies and two principles: The exigencies are thermodynamic stability and high density. The principles are that information is information and that information conveyance requires a code. Thermodynamic Stability

Abstract not found