Extensions and evaluations of adaptive processing of structured information using artifical neural networks

The application of Artificial Neural Networks has traditionally been restricted to fixed size data and data sequences. However, there are a large number of applications which are more appropriately represented in the form of graphs. Such applications include learning problems from the area of molecular chemistry, software engineering, artificial intelligence, image and document processing, and numerous others. The inability of conventional Artificial Neural Networks to encode this kind of data has motivated for research in this field.