Conclusions

Abstract

The contribution tree shown in Figure 1.3 outlines the contributions of this monograph. To be more specific, this study has achieved the following:

• 1. Set up a formal conceptual structure (ontology) for multi-level adaptation of discrete time artificial neural network systems .

• 2. Developed methods to decide the minimum delay for signal propagation through the network thus ensuring state consistency.

• 3. Introduced an activity-based structure level adaptation paradigm as a general guideline for the structure level adaptation of neural network systems. In particular, I proposed a general procedure for designing adaptable structure neural networks and listed the criteria for guiding the firing of various structural operators.