Improving Tools in Artificial Intelligence

Abstract

The historical origin of the Artificial Intelligence (AI) is usually established in the Dartmouth Conference, of 1956. But we can find many more arcane origins [1]. Also, we can consider, in more recent times, very great thinkers, as Janos Neumann (then, John von Neumann, arrived in USA), Norbert Wiener, Alan Mathison Turing, or Lofti Zadeh, for instance [12, 14]. Frequently AI requires Logic. But its Classical version shows too many insufficiencies. So, it was necessary to introduce more sophisticated tools, as Fuzzy Logic, Modal Logic, Non-Monotonic Logic and so on [1, 2]. Among the things that AI needs to represent are categories, objects, properties, relations between objects, situations, states, time, events, causes and effects, knowledge about knowledge, and so on. The problems in AI can be classified in two general types [3, 5], search problems and representation problems. On this last "peak", there exist different ways to reach their summit. So, we have [4] Logics, Rules, Frames, Associative Nets, Scripts, and so on, many times connected among them. We attempt, in this paper, a panoramic vision of the scope of application of such representation methods in AI. The two more disputable questions of both modern philosophy of mind and AI will be perhaps the Turing Test and the Chinese Room Argument. To elucidate these very difficult questions, see our final note.