Abstract
New algorithms for ultrafast (single iteration) learning in feedforward neural networks are developed. In addition, a methodology to determine the confidence limits of results predicted by neural network models is formulated. This methodology also consistently combines experimental data (e.g., sensor measurements) with model-predicted results. Our goal is to obtain best estimates for the ne twork model parameters, and to drastically reduce the uncertainties underlying decision processes based on learning. Preliminary results of applying the approach to seismic analysis are presented. These results show remarkable promise for petroleum reservo ir characterization.
Original language | English |
---|---|
Title of host publication | Advances in Neural Networks and Applications |
Publisher | World Scientific and Engineering Academy and Society |
Pages | 427-432 |
Number of pages | 6 |
ISBN (Print) | 9608052262 |
State | Published - 2001 |
Keywords
- Petroleum reservoir characterization
- Seismic analysis
- Ultrafast learning
- Uncertainty reduction
- Virtual layer