Abstract
Analysis of the error surfaces of feed-forward neural networks is complicated by the high dimensionality of the weight space. Visualisation over one- and two-dimensional slices, and Monte Carlo analysis of stationary points can produce misleading results. We show that, in some situations, important features of the error surface can only be visualised by considering the error over non-planar manifolds of weight space. We also show that Monte Carlo simulations can depend critically upon the random step size chosen. The relationship can reveal key properties of the local structure of the error surface.
| Original language | English |
|---|---|
| Title of host publication | Proceedings of the Sixth Australian Conference on Neural Networks |
| Editors | M Charles, C Latimer |
| Place of Publication | Sydney |
| Publisher | The University of Sydney |
| Pages | 197-200 |
| Number of pages | 4 |
| ISBN (Print) | 0909391033 |
| Publication status | Published - 1995 |
Fingerprint Dive into the research topics of 'The structure of neural network error surfaces'. Together they form a unique fingerprint.
Cite this
Hamey, L. (1995). The structure of neural network error surfaces. In M. Charles, & C. Latimer (Eds.), Proceedings of the Sixth Australian Conference on Neural Networks (pp. 197-200). Sydney: The University of Sydney.