Connectionists Modelling in Letter and Words Recognition Essay

Connectionists Modelling in Letter and Words Recognition - Essay Example These simple units can represent neuron and the connections can represent synapse in the neural network. Biological Activism: Neural network of connectionist modeling suggests that the study of mental activity is the study of neural systems. This links connectionism to neuroscience, and models involve varying degrees of biological realism. The biological aspects of natural neural systems are incorporated in connectionist model for better understanding / biological reality. Learning: Learning is an important aspect of connectionist modelling. Many sophisticated learning procedures for neural networks have evolved, modifying the connection weights. Mathematical formulas are used to determine the change in weights when given sets of data consisting of activation vectors for some subset of the neural units. Parallel Distributed Processing: It is a neural network approach emphasizing the parallel nature of neural processing and the distributed nature of neural representations. It provides a general mathematical framework for researchers to operate in. The framework involved eight major aspects: These aspects are now the foundation for almost all connectionist models. It is assumed that all cognitive processes are explained by neural firing and communication. According to this view there is no room for rational thinking or emotion. Discovery of methods for training multilayer networks is the ... 'Activation rule for combining inputs to a unit to determine its new activation'- represented by a function on the current activation and propagation. 'Learning rule for modifying connections based on experience'- represented by a change in the weights based on any number of variables. 'Environment which provides the system with experience'- represented by sets of activation vectors for some subset of the units. These aspects are now the foundation for almost all connectionist models. It is assumed that all cognitive processes are explained by neural firing and communication. According to this view there is no room for rational thinking or emotion. Discovery of methods for training multilayer networks is the major turning point in connectionist modeling. With this discovery, connectionist models not only have the computational power to answer those questions interesting to cognitive science, but also have a method of learning how to answer those questions. Thus, there is an explicit distinction between network architectures and the learning rules used to train them within new connectionism. By understanding the different types of architectures and learning rules, researchers are in a position to choose the appropriate type of network to solve specific problems. For example, if one wanted to solve a pattern recognition problem that was linearly separable, then an integration device network would be appropriate. If the problem was linearly inseparable, however, then the value unit architecture would be more appropriate. Advantages of Connectionist modeling: Connectionist modelling engage in "low level" modeling, trying to ensure that their models resemble neurological structures. Connectionist modelling focus on