10. Future Trends

Neural network has received a great deal of attention over the past decade among scientists and engineers in all the disciplines. The term "neural network" resulted from artificial intelligence, that aims at understanding and modeling brain behavior. A neural network can be considered as an expert computing system comprising of a number of simple, highly interconnected nodes or processing elements, in which information is processed by their dynamic state response to external inputs. The goal of a neural network is map a set of input patterns onto a corresponding set of output patterns. This mapping is achieved by neural network system by first learning from a series of past examples defining sets of input and output correspondences. The network then applies what it has learned to a new input to predict the appropriate output.²⁸

A typical neural network architecture consists of various nodes that are interconnected via links that are used that are use to transfer the information. The neural network has an input layer, at least one (normally 1 to 3) hidden layer, and an output layer. The whole network can be viewed as a "black box" into which specific inputs (each corresponding to a specific node) are sent. This information is processed through the interconnections between nodes, but the entire processing step is hidden from the user. Finally, an output is given from the nodes on the output layer.²⁸

“Knowledge based” control of fermentation processes is an exciting area of research devoted to solving control problems associated with the uncertainties and nonquantitative nature of biochemical systems. The use of fuzzy logic is most attractive for fermentation processes because of the poorly understood dynamics but with a good amount of process information accumulated as experienced knowledge of a human operator.²⁹