Compare RBF and MLP:
RBF |
MLP |
1. RBFN is a ingle hidden layer. |
1. MLP is a multiple hidden layer. |
--- |
--- |
2. In RBF hidden layer computation nodes are different from output nodes. |
2. MLP follows the common computational model in hidden as well as output. |
3. In RBF hidden layer is non-linear and output layer is linear. |
3. In MLP hidden layer and output layer is linear. |
4. The argument of RBF activation function computes Euclidean norm between input vector and centre. |
4. Each hidden unit computes the inner product of input vector and synaptic vector. |
5. Exponentially decaying local characteristics. |
5. Global approximation to non-linear input - output mapping. |
6. RBFN is fully connected. |
6. MLP can be partially connected. |
7. In RBFN, the hidden nodes operate differently i.e. they have different models. |
7. In MLP, the hidden nodes share a common model not necessary the same activation function. |
8. In RBF network we take differece of input vector and weight vector |
8. In MLP network we take product of input vector and weight vector. |
9. In RBF training of 1 layer at a time. |
9. In MLP training of all layer simultaneously. |
10. RBFN does faster training process. |
10. MLP is slower in training process. |
11. RBFN is slow when practically used. |
11. MLP is faster when practically used. |
(ii) How do you achieve fast learning in ART 2 network.
- Increasingly popular neural networks with diverse applications and the main rival to the multi layered perceptron.
- Input layer is simple a fan out and does no processing.
- The second or hidden layer performs a non-linear mapping from the input space into a higher dimensional space in which the patterns become linearly separable.
- Weights of the hidden layer corresponds to cluster center, output function is usually Gaussian.
- The hidden units provide a set of function that constitute an arbitrary "basis" for the input patterns when they are expanded into the hidden space. These functions are called Radial Basis Functions.
- In most application the hidden space is of high dimensionality.
- The output layer performs a simple weighted sum with a linear output.
- A mathematical justification for the rational of a nonlinear transformation followed by a linear transformation may be traced back to an early paper by cover.
- The underlying justification is found in cover's theorem which states that "A complex pattern classification problem cast in high dimensional space non-linearly is more likely to be linearly separable that in a low dimensional space.
- This is the reason for making the dimension of the hidden space in an RBF network high.
- We know that once we have linearly separable patterns, the classification problem is easy to solve.
- It is easy to have an RBF network perform classification. We simply need to have a an output function yk(x) for each class k with appropriate targets and when the network is trained it will automatically classify new patterns.
- The most commonly used Radial Basis Function is a Gaussian function.
- In a RBF network r is the distance from the cluster center.
- The distance measure from the cluster center is usually the Euclidean distance.
- For each neuron in the hidden layer the weights represent the co-ordinates the center of the cluster.