- Published on
KembaraXtra-Psychology – Back-Propagation Algorithm
The back-propagation algorithm is a learning method used in parallel distributed processing systems. A multilayer neural network adjusts its internal connection weights by comparing its output with a desired target output, calculating the error, and propagating corrections backward through the layers using gradient descent. This iterative process allows the system to learn complex input–output relationships and became influential through work by David Rumelhart and colleagues in the 1980s, building on earlier ideas by Frank Rosenblatt.
0 Comments