The problem

The goal of this research is to build an integrated environment to serve as a possible assistant in the educational process. Getting useful information for the teaching process is very difficult when dealing with unstructured knowledge.

This goal is divided into subgoals or subproblems, which take a modular form. The complete Subsymbolic Knowledge Extraction Environment (SSKEE) system contains three building blocks: the neural network module, the knowledge extraction module and the user interface module.

The subproblems and goals are:

• Finding and usage of an energy function for supervised learning on feedforward networks, based on restrictions, that is used for the neural network module, as will be seen in part I of the thesis.
• Finding optimizations for the neural network module by using parallelism. In the context of the ANN implementation, I discuss the parallel aspects of a ANN, as well as the advantages of a parallel implementation of a ANN. Then I design a correspondence between the different levels of parallelism, namely, coarse, medium and fine grain parallelism and the ANN functions and structure. The optimizations made can be seen in part II of the thesis.
• An other subgoal is designing of a subsymbolic knowledge extraction environment. As the teaching process requires only symbolic knowledge, I believe the knowledge extraction tool to be a chance for teachers to significantly improve their teaching materials and/or style by combining the symbolic knowledge of the domain theory with the rules extracted from the empirical sub-symbolic knowledge stored in ANNs trained on examples. This is the subject of the part III of the thesis.
• The last goal is to design a user interface for the whole system. The third module, the user interface module is the one to make the link to the end user, in this case, the professor/teacher. The development of this interface is discussed throughout the thesis, as each of the modules have an independent interface.

The whole system has to add up to a Neural Network's Sub-Symbolic Knowledge Eliciting Environment for possible Education Process Assistance from a Time-Series prediction parallel Neural Network.