Home > April, 2009
If so, could an evolving neural network solve any problem given arbitrarily large training inputs and processing time?
Well, you would have to build the network to be Turing-complete.
And yes, given an infinite amount of time and an infinitely powerful computing device they can solve any problem, as long as it is asked properly.
powered by Yahoo Answers
A neural network computer has a network of something, what are the things that are networked? Are they individual processors? Some other form of logic?
I get that part, but what are the "neurons"? Is that just a general name for a decision making element?
Network of networks. At the lowest level are the actual "neurons" - implemented in models - which form an interconnected decision-making matrix. It's more a reflection of logic connections than specifying physical components. It's not a network of computers or processors. It is a network of neurons. These neurons (the net) may indeed be spread out over the network or amongst CPUs or servers or workstations, but that's merely an implementation description, not how they work.
powered by Yahoo Answers
Try to run a search on google. I ran a search on good with key words "artifical intelligence", prediction of pavement distresses and found several.
Here is one for sample.
http://gulliver.trb.org/publications/circulars/ec012.pdf
Other search results..
http://www.google.com/search?hl=en&lr=&q=artificial+intelligence%2C+prediction+of+pavement+distresses
powered by Yahoo Answers
Hello, I'm trying to get my head around this neural network stuff. So a neuron's output is the sigmoid function applied to the total sum of all weights added together? So then where does the treshold fit in, and what is a bias? Am I missing anything? Thankyou. Have a good day.
The terminology is a bit confusing. In the common type of feed-forward back-propagation artificial neural network using a sigmoid transfer function the sigmoid function is applied to the sum of the weighted inputs i.e. to the sum of w_i * v_i where w_i is the weight associated with the input i and v_i is the current value of input i. Applying the sigmoid function to just the sum of the weights would mean that the neuron's output didn't change when the inputs changed, which would be useless.
The bias is an offset value applied to this sum - in effect, it is the weight of an input whose value v_bias is always 1. The bias shifts the centre of the sigmoid function away from the origin. The term originates in signal processing filter design, I think.
Threshold is used for basic perceptrons, neurons whose outputs take discrete values (e.g. where the output is either 0 or 1). If the value of the sigmoid function crosses the threshold, then the neuron's output switches from one value to another. I get the impression that switching perceptrons are rarely used in data interpretation these days except in a teaching context. Setting the neuron's output equal to the value of the sigmoid function, instead of using the sigmoid function value as a trigger to switch between different set output values, is more powerful. If your network's output values must be discrete (e.g. you are encoding a logic circuit as an exercise) then using a switching perceptron might be appropriate, at least in the final layer.
The terminology I'd recommend is feed-forward back-propagation artifical neural network. Feed-forward identifies how a network with a given set of connection weights and biases processes the input values to produce the output values. Back-propagation identifies how the network is trained using sets of known input/output data to adjust the values of the weights and biases. Again, the bias is simply a weight applied to an input whose value is always one. Once trained, an artificial neural network can be run in feed-forward mode only.
powered by Yahoo Answers
I am a Science student and I want to know what are neural networks. Can u please explain me?
Neural Networks
A neural network consists of many simple processing units that are connected by communication channels. Much of the inspiration for the field of neural networks came from the desire to perform artificial systems capable of sophisticated, perhaps intelligent computations similar to those of the human brain.
Neural networks usually learn from examples and exhibit some capability for generalization beyond the data used for training. They are able to approximate highly nonlinear functional relationships in data sets.
Figure 1: A neuron within a neural network.
The smallest part of a neural network is one single neuron as shown in Figure 1. It takes a set of individual inputs and determines (through the learning algorithm) the optimal connection weights that are appropriate to each input. Next, the neuron aggregates these weighted values to a single value
An activation function is then applied to the aggregated weighted value to produce an individual output
for the specific neuron. A typical activation function is the logistic distribution function
The aim of a neural network is to explain the outputs by the input variables . More exactly, we want to find functions such that explains the output variable .
A neural network with one hidden layer (single hidden layer) consists of neurons of three basic types:
The input neurons collect the external information and send it to the layer of hidden units.
The hidden neurons aggregate the information and send it to the output neuron(s).
The output neurons contain the aggregated information passed through the activation function.
powered by Yahoo Answers
Lecture Series on Artificial Intelligence by Prof.Sudeshna Sarkar and Prof.Anupam Basu, Department of Computer Science and Engineering,IIT, Kharagpur . For more details on NPTEL visit nptel.iitm.ac.in.
Duration : 1:0:46
how far the research is done in ayurvedic medicine applications using artificial neural networks?
That's quite an impressive interest. I have heard very good things about artificial neural medicine in the past few months (matter of fact, heard something about it on the news the other night).
powered by Yahoo Answers
what is the problem for which neural network is a solution .what is the technique adopted for image processing before invention of neural network.is neural network just a way of programming where processor can be made to learn to solve the process.will neural network get trained even during the testing.can face recognition be done by storing a database of pictures and comparing the test face based on pixel by pixel information with the stored database of faces.what are the ways of implementing face recognition without using neural networks.
- The very important reason to move towards Neural Network rather than pure Image Processing algorithms may rely on their error and variation tolerance.
- Neural Networks easily adopt to the environment while pure mathematics in image processing is rigid.
- Sometimes formulating a phenomenon is too difficult or even impossible, while Neural Networks do the Rule(Information) Extraction automatically.
Q. what is the technique adopted for image processing before invention of neural network?
A. You can’t call a single method adopted. You should know that before inventing Neural Networks, people had many difficulties in their pattern recognition. That was mostly because they had to invent a new way of processing for every case they encountered.
Q.will neural network get trained even during the testing ?
A. In Backpropagation, it depends on the way you implement it. While in those which are on unsupervised techiniques, there is actually no testing; whatever you show the network goes for learning.
Q.can face recognition be done by storing a database of pictures and comparing the test face based on pixel by pixel information with the stored database of faces?
A. Sure. That’s the way a BPNN works.
I refer you to the following link, if you want to use a BPNN, for training, testing, recognition, etc. :
1) go to http://www.MehranHoodeh.com
2) Click on “Experiences”
3) Click on “Neural Networks”
4) Download “MPE”.
5) Extract the compressed file.
6) Run “MPE.exe”
7) Go to “Help -> Offline Help” menu option.
Have a read through its documents.
9) Run samples. (Start with XOR)
10) Any more Questions? Then email me.
powered by Yahoo Answers
I want some applications of neural network in details?
Several years ago, I developed a set of 4 neural networks for a manufacturing company, which predicted their industry's microeconomic production for 3, 6, 9 and 12 months out. The models accepted as inputs: previous values of production, as well as several other economic variables. The output was presented as a point forecast and expected 90% prediction intervals.
powered by Yahoo Answers
Learn Neural Net Programming: www.heatonresearch.com Introduction to Neural Networks with C# is a course that introduces the C# programmer to the world of Neural Networks and Artificial Intelligence. Neural network architectures, such as the feedforward, Hopfield, and self-organizing map architectures will be presented. Training techniques, such as backpropagation, genetic algorithms and simulated annealing are also introduced. Practical examples are given for each neural network. Examples …
Duration : 0:10:1