Abstract | ||
---|---|---|
The artificial neural network with one hidden unit and the input units connected to the output unit is considered. It is proven that the error surface of this network for the patterns of the XOR problem has minimum values with zero error and that all other stationary points of the error surface are saddlepoints. Also, the volume of the regions in weight space with saddlepoints is zero, hence training this network on the four patterns of the XOR problem using, e.g., backpropagation with momentum, the correct solution with error zero will be reached in the limit with probability one. |
Year | DOI | Venue |
---|---|---|
1996 | 10.1162/neco.1996.8.6.1301 | Neural Computation |
Keywords | Field | DocType |
error surface,zero error,minimum value,input unit,output unit,error zero,xor problem,simplest xor network,hidden unit,artificial neural network,correct solution,global minimum,local minima,backpropagation,saddle point | Artificial intelligence,Momentum,Artificial neural network,Weight space,Error surface,Network on,Mathematical optimization,Algorithm,Maxima and minima,Stationary point,Backpropagation,Mathematics,Machine learning | Journal |
Volume | Issue | ISSN |
8 | 6 | 0899-7667 |
Citations | PageRank | References |
13 | 3.14 | 1 |
Authors | ||
2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Ida G. Sprinkhuizen-kuyper | 1 | 84 | 13.83 |
Egbert J. W. Boers | 2 | 122 | 20.73 |