Input neurons

The architecture of a backpropagation neuronal network is comparatively simple. The network consists of different neurone layers. The layer connected to the network input is called input layer while the layer at the network output is called the output layer. The different layers between input and output are named hidden layers. The number of neurones in the layers is determined by the developer of the network. Networks used for classification have commonly as much input neurones as there are features and as much output neurones as there are classes to be separated. 

An appropriate perceptron model for this problem is one that has two input neurons (corresponding to inputs X and X2) and one output neuron, whose value... 

In the more general case when there are N input neurons, the space of the Xi s... 

We have seen that the output neuron in a binary-threshold perceptron without hidden layers can only specify on which side of a particular hyperplane the input lies. Its decision region consists simply of a half-plane bounded by a hyperplane. If one hidden layer is added, however, the neurons in the hidden layer effectively take an intersection (i.e. a Boolean AND operation) of the half-planes formed by the input neurons and can thus form arbitrary (possible unbounded) convex regions. ... 

The number of sides of the convex regions is equal to the miinber of half-planes whose intersection formed the decision region, and is thus bounded by the number of input neurons. 

Fig. 6.18. Schematic representation of a multilayer perceptron with two input neurons, three hidden neurons (with sigmoid transfer functions), and two output neurons (with sigmoid transfer functions, too)...

Neural Nets (NNs) relate a set of input neurons with an output neuron (providing the prediction label of a data point) by a network of layers of neurons in the interior. They are certainly among the most frequently used Machine Learning methods in the field [148] and allow for a high degree of customization since the architecture of the network itself is part of the parameters the user may define. 

The previously scaled atomic spectrum of a standard (technically, it is called a training pattern) enters the net throughout the input layer (a variable per node). Thus, an input neuron receives, simply, the information corresponding to a predictor variable and transmits it to each neuron of the hidden layer (see Figures 5.1, 5.3 and 5.4). The overall net input at neuron7 of the hidden layer is given by eqn (5.3), which corresponds to eqn (5.1) above ... 

Figure 7.7 Diagram of an artificial neuron connected to three input neurons and a bias unit. (From Manallack, D.T. and Livingstone, D.J., Med. Chem. Res., 2, 181-190, 1992. With permission.)...

The dendrites represent all the processes of the cell body except for the specialized axonal process (axon). They are usually numerous and serve to increase the surface area of the neuron available for receiving synaptic input. Neurons will have one or more main dendrites that successively branch and arborize to form many smaller processes. 

So, the basic neuron can be seen as having two operations, summation and thresholding, as illustrated in Figure 2.5. Other forms of thresholding and, indeed, other transfer functions are commonly used in neural network modeling some of these will be discussed later. For input neurons, the transfer function is typically assumed to be unity, i.e., the input signal is passed through without modification as output to the next layer F(x) = 1.0. 

Figure 5.13 Layers of units and connection links in an artificial neuronal network. Ij-ij.- input neurons, h-hj hidden neurons, b, b exit and output bias neurons, w,2hi weight of transmission, ij-h, x.-Xji input process variables and y output...

Isaac JTR, Crair MC, Nicoll RA, Malenka RC (1997) Silent synapses during development of thalamocortical inputs. Neuron 18 269-280. 

The solution of the exact interpolating RBF mapping passes through every data point (x , y ). In the presence of noise, the exact solution of the interpolation problem is typically a function oscillating between the given data points. An additional problem with the exact interpolation procedure is that the number of basis functions is equal to the number of data points, so calculating the inverse of the N x N matrix becomes intractable in practice. The interpretation of the RBF method as an artificial neural network consists of three layers a layer of input neurons feeding the feature vectors into the network a hidden layer of RBF... 

In parallel to the SUBSTRUCT analysis, a three-layered artificial neural network was trained to classify CNS-i- and CNS- compounds. As mentioned previously, for any classification the descriptor selection is a cmcial step. Chose and Crippen published a compilation of 120 different descriptors, which were used to calculate AlogP values as weU as drug-likeness [53, 54]. Here, 92 of the 120 descriptors and the same datasets for training and tests as for the SUBSTRUCT algorithm were used. The network consisted of 92 input neurons, five hidden neurons, and one output neuron. 

Figure 6 Principal structure of a backpropagation network a) Information processing of a single neuron b) Exemplaric structure of a network with two input neurons, one hidden layer of three neurons and an output layer consisting of one output neuron.

Using the original piincijjal compwnent scores as inputs, the best architecture consist of a three layer network with 23 input neurons, 10 neurons in the hidden layer and one neuron in the output layer. Considering RPC scores as inputs, the best architectures were achieved with almost the same number of hidden neurons. The hidden neurons consist of 9 and 10 neurons respectively. Training was carried out for a maximum 10000 iterations. Selection of the network was performed at maximum correlation coefficient (R) and 95% confidence limit. 

In the case of two input neurons, x and X2, for the following model for the separation plane... 

The network consists of two input neurons for presentation of the two v-values as well as four output neurons, y, which represent the four classes (cf Figure 8.15). In addition, a hidden layer was added with up to 20 neurons and the intercepts of the surfaces are modeled by bias neurons to both the hidden and output layers. The transfer function in the neurons of the hidden layer was of sigmoid type, and aggregation of the neurons in the output layer was carried out by calculating the normalized exponentials (softmax criterion). 

---