An introduction to simulated neural networks

Neural networks have the potential to offer an exciting advance over traditional analytical methods for solving heretofore intractable problems. Despite providing very little in terms of causal understanding, they can more accurately replicate some estimation processes. One area of potential application of neural networks in resource management is the rating of the aesthetic value of natural scenes. In this example, a neural network was generated which had actually "learned" how to replicate the aesthetics-scoring mechanisms that took place within the minds of our human subjects. While some success has been encountered using neural networks to address a range of problems, there are some potential pitfalls. Unless one is absolutely confident of the comprehensiveness of the training data, the results of such analyses could, logically, be quite misleading. Hence it is easy to become overly enthusiastic about mistaken interpretation of experimental, neural-network-based results.