Autoencoders actually reconstructs the original input and also it helps in dimensionality reduction as the number of hidden neurons is less compared to the number of input neurons. My question is how output values are generated from the hidden neuron values? WHAT IS THE MATHEMATICAL FORMULA THAT IS USED TO CALCULATE THE FINAL OUTPUT VALUES(starting from input to hidden and hidden to output). PLEASE anyone help me with this. I have tried mathematically,but I am not getting the output as same as the input values.
Asked
Active
Viewed 76 times
1 Answers
0
There isn't a set, single, way to feed-forward neural nets - it's a general technique. One popular thing to do is logistic(W*In), where W*In is the dot product of the node's weights and the input nodes' activations, and logistic(x) = 1/(1+e^-x). There are many, many subtleties to applying this method, and the "meat" of the technique is how you determine / train the weights W for each node. I recommend getting a good text on machine learning / neural networks, perhaps - (even if it's not specifically talking about autoencoders, the general techniques used for multi-layer nets will be similar): http://www.amazon.com/Pattern-Recognition-Learning-Information-Statistics/dp/0387310738/ref=zg_bs_3894_3
BadZen
- 4,083
- 2
- 25
- 48