Artificial Intelligence in Computer Science

1080 Words5 Pages
An important field in computer science today is artificial intelligence. The novel approaches that computer scientists use in this field are looked to for answers to many of the problems that have not been solved through traditional approaches to software engineering thus far. One of the concepts studied and implemented for a variety of tasks in artificial intelligence today is neural networks; they have proven successful in offering an approach to some problems in the field, but they also have some failings.

Traditional neural networks, which “learn” by changing the values, or weights, contained at nodes in a directed graph, suffer from several issues that make actually applying them to a given problem difficult and unwieldy. They require large amounts and frequent application of training—the material with which their knowledge banks are kept accurate—which makes them difficult to maintain. Many neural network systems suffer from a type of information overload in which they may lose data after being trained on a large dataset, which may not be desirable, as it could damage the reliability of the produced results.

The various deficiencies of traditional neural networks, coupled with their encouraging success in some areas, have prompted research into alternate network models. A new type of neural network proposed by Dr. Anthony Beavers, the dynamic associative network (DAN), offers an answer to some of traditional neural networks’ problems. Instead of altering the weights of nodes during training, DANs simply add new interconnected nodes as they are needed. They need not be trained continuously to retain their correct behavior, unlike traditional neural networks, and they are not data-lossy. Recent research suggests that ...

... middle of paper ...

...feature detectors will be in a particular partially activated state, which is called a signature. This signature will be matched against all the other unique signatures constructed from the InPhO database, and using the rich partial-match properties of dynamic associative networks, determine which elements of the ontology have the most in common with the Document. This set of elements will further be analyzed to locate the appropriate position of the Document within the hierarchy of philosophy.

The project as proposed has no significant economic, environmental, health and safety, or political ramifications. One of the goals is to minimize the influence of (and number of hours required from) the philosophy professionals that currently help to curate the taxonomy, which could have social consequences, as their dedication of time and effort would no longer be needed.
Open Document