ANNs are powerful computational tools that can be used for classification, pattern recognition, empirical modeling, and many other tasks. Even though most of these tasks can also be performed by conventional statistical or mathematical methods, ANNs often provide a more effective way to deal with problems that are difficult, if not intractable, for traditional computation. In fact, while traditional computation is based on the a priori selection of suitable functions or algorithms, ANNs are able to adjust their inner structures to provide optimal solutions, given enough data and a proper initialization. Thus, if appropriate inputs are applied to an ANN, it can acquire knowledge from the environment, mimicking the functioning of a brain, and users can later recall this knowledge.
Was this article helpful?