EP Downhill Walkers Explore a Landscape in Weight Space of a Fully-connected Neural Network Model: Why does Hebbian Peak Resist to be Found?
نویسنده
چکیده
Using various versions of evolutionary computation, we explore fitness landscape defined on continuous synaptic weight space of a fully connected neural network model of associative memory. Thus far, we have found many different solutions that give a function of associative memory to a network. However, the Hebbian peak, one of the most familiar solutions, has never been located in these explorations. To address this issue, we study shape of peaks on the fitness landscape by reverse hill-climbing walk taking advantage of the evolutionary programming. We guess why does the Hebbian peak resist to be found.
منابع مشابه
Estimation of Network Reliability for a Fully Connected Network with Unreliable Nodes and Unreliable Edges using Neuro Optimization
In this paper it is tried to estimate the reliability of a fully connected network of some unreliable nodes and unreliable connections (edges) between them. The proliferation of electronic messaging has been witnessed during the last few years. The acute problem of node failure and connection failure is frequently encountered in communication through various types of networks. We know that a ne...
متن کاملA Convolutional Neural Network based on Adaptive Pooling for Classification of Noisy Images
Convolutional neural network is one of the effective methods for classifying images that performs learning using convolutional, pooling and fully-connected layers. All kinds of noise disrupt the operation of this network. Noise images reduce classification accuracy and increase convolutional neural network training time. Noise is an unwanted signal that destroys the original signal. Noise chang...
متن کاملEnhancing Efficiency of Neural Network Model in Prediction of Firms Financial Crisis Using Input Space Dimension Reduction Techniques
The main focus in this study is on data pre-processing, reduction in number of inputs or input space size reduction the purpose of which is the justified generalization of data set in smaller dimensions without losing the most significant data. In case the input space is large, the most important input variables can be identified from which insignificant variables are eliminated, or a variable ...
متن کاملWhat does the Landscape of a Hopfield Associative Memory Look Like?
We apply evolutionary computations to the Hop eld's neural network model of associative memory. In the model, some of the appropriate con gurations of synaptic weights give the network a function of associative memory. One of our goals is to obtain the distribution of these con gurations in the synaptic weight space. In other words, our aim is to learn a geometry of a tness landscape de ned on ...
متن کاملPersian Phone Recognition Using Acoustic Landmarks and Neural Network-based variability compensation methods
Speech recognition is a subfield of artificial intelligence that develops technologies to convert speech utterance into transcription. So far, various methods such as hidden Markov models and artificial neural networks have been used to develop speech recognition systems. In most of these systems, the speech signal frames are processed uniformly, while the information is not evenly distributed ...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 2001