4. Implementation
To be able to implement the described methods and techniques, the order of importance of the list below must be clear:
• The foundation is the most important;
• Methods and techniques are less important;
• Tools are the least important.
Methods and techniques
Since the described methods and techniques have a fundamental approach, the most obvious methods and techniques – like NLP and ontology – might be insufficient to implement them.
Examples:
• Ontology is designed to represent a logical structure – rather than sentences in natural language;
• NLP techniques are designed to parse sentences – rather than to build a knowledge structure from them or to represent a logical structure;
• Both NLP and ontology are not designed to be compatible: formal language versus natural language, unlike the above described reasoning and knowledge representation techniques.
Tools
In the same way, the most obvious tools – like Prolog – might be insufficient to implement the described fundamental approach, because e.g. like ontology and Prolog are designed with logic in mind – programmation en logique – rather than representing sentences in natural language.
Covering multiple disciplines
The system described above combines separate disciplines in the field of AI, in which the methods, techniques and tools are also designed to only cover one specific discipline. Therefore, it will be hard to combine those methods, techniques and tools, in order to implement the described system.
Contribution to science
The described system combines separate disciplines in the field of AI: Parsing (NLP), reasoning (ontology), semantics and writing new knowledge in grammatically correct sentences.