Learning algorithm for multi-layered networks, Computer Engineering

Learning algorithm for multi-layered networks:

Furthermore details we see that  if S is too high, the contribution from wi * xi is reduced. It means that t(E) - o(E) is multiplied by xi after then if xi is a big value as positive or negative so the change to the weight will be greater. Here to get a better feel for why this direction  correction works so it's a good idea to do some simple calculations by hand. 

Here η simply controls how far the correction should go at one time that is usually set to be a fairly low value, e.g., 0.1. However the weight learning problem can be seen as finding the global minimum error which calculated as the proportion of mis-categorised training examples or over a space when all the input values can vary. Means it is possible to move too far in a direction and improve one particular weight to the detriment of the overall sum: whereas the sum may work for the training example being looked at and it may no longer be a good value for categorising all the examples correctly. Conversely for this reason here η restricts the amount of movement possible. Whether large movement is in reality required for a weight then this will happen over a series of iterations by the example set. But there sometimes η is set to decay as the number of that iterations through the entire set of training examples increases it means, can move more slowly towards the global minimum in order not to overshoot in one direction.

However this kind of gradient descent is at the heart of the learning algorithm for multi-layered networks that are discussed in the next lecture. 

Further Perceptrons with step functions have limited abilities where it comes to the range of concepts that can be learned and as discussed in a later section. The other one way to improve matters is to replace the threshold function into a linear unit through which the network outputs a real value, before than a 1 or -1. Conversely this enables us to use another rule that called the delta rule where it is also based on gradient descent.

Posted Date: 1/11/2013 7:02:52 AM | Location : United States

Related Discussions:- Learning algorithm for multi-layered networks, Assignment Help, Ask Question on Learning algorithm for multi-layered networks, Get Answer, Expert's Help, Learning algorithm for multi-layered networks Discussions

Write discussion on Learning algorithm for multi-layered networks
Your posts are moderated
Related Questions
Customer arrivals at a gas station can be characterized by exponential distribution with mean of 10 minutes. The amount of time they take to pump gas can be characterized by an ex

What are the special unit related fields and methods?   The most significant method (in fact pseudo method) related to units is get_enclosing_unit().  The mostly used field in

Q. Explain about Wide Area Network? Wide Area Network (WAN) usually refers to a network, which covers a large geographical area as well as use communications subnets (circuits)

What is recursion?  A recursive procedure should have two properties.  What are they? Recursion means function call itself repeatedly.         Properties : (1) There

What are Attributes?  Attributes are declarative tags in code that insert additional metadata into an assembly. There exist two types of attributes in the .NET Framework: Pred

Multiple Instruction and Multiple Data stream (MIMD) In this association, multiple control units and multiple processing elements are prepared as in MISD. But the discrepancy i

What is indexing? Specific fields shown on each scanned document are provided to our organization to make the systematic arrangement of your records. This process is designed t

Explain the working of Interpreter? The Interpreter is a program that executes instructions written in a high-level language. There are two manners to run programs written in a

Q. Register-to-register operands in RISC? Register-to-register operands: In RISC machines operation which access memories are LOAD and STORE. All other operands are kept in reg