Follow project on Twitter

AI and NLP – a fundamental approach

(the downfall of the theory of evolution as the assumed origin of intelligence and language)

Download the fundamental document (updated: June 25, 2016):
AI and NLP – a fundamental approach.pdf


Around the year 1956, the field of Artificial Intelligence (AI) and knowledge technology was started. However, there are four reasons to believe that the current approach to AI and knowledge technology has a fundamental problem:

1. Intelligence and language are natural phenomena. Natural phenomena obey laws of nature. And laws of nature are investigated using (basic or) fundamental science. However, the field AI and knowledge technology is researched using (behavioral or) cognitive science. This cognitive approach delivers a simulation of behavior (similar to a flight simulator), while a fundamental approach would deliver an artificial implementation of natural intelligence (similar to an airplane);
Fundamental science versus cognitive science
2. A science has a foundation in nature, which leads to generic solutions. But due to its cognitive approach, the field of AI and knowledge technology has no foundation in nature – nor a definition based on nature – 60 years after its start. Without foundation, this field is baseless. And being baseless, this field is limited to engineer specific solutions to specific problems, while a science delivers generic solutions;

3. As a consequence, knowledge technology is based on applying smart algorithms to keywords, by which non-keywords are ignored. Non-keywords provide information to our brain about the structure of the sentence. But by ignoring this structure provided by nature, the field of knowledge technology got stuck with “bags of keywords” and unstructured texts.

4. Moreover, a science integrates its disciplines. However, the field of AI and knowledge technology fails to integrate (automated) reasoning and natural language. In other words, this field has a blind spot:
• Reasoners (like Prolog) are able to reason, but their results – derived knowledge – can't be expressed in readable and automatically constructed sentences. Even Watson of IBM has only a keyword as output. The Jeopardy game version: “What is {noun}?” or “Who is {proper noun}?”;
• Chatbots and Virtual (Personal) Assistants may well produce understandable sentences, but they are unable to reason logically. Moreover, they are only able to select a human-written sentence, in which they may fill-in a user-written keyword;
• Controlled Natural Language (CNL) reasoners are very limited in integrating both disciplines. They are limited to sentences with verb “is/are”, and don't accept words like definite article “the”, conjunction “or”, possessive verb “has/have” and past tense verbs “was/were” and “had”.