Sixty plus years ago, Alan Turing had formulated the ‘Imitation Game’. It was alternative way of asking and describing the question, “Can Machines think?” Through his description he had defined ‘machine’ and ‘think’ and concluded that there is plenty ahead that needs to be done. It was the time when digital computers were not realized yet. It was back then, when machines took thirty seconds to compute the addition of two five digit numbers. Some of the thoughts expressed are still relevant, off course in current context and progress.
Machine learning and Artificial Intelligence has progressed for writing a better ‘book of rules’ to train the machine. The question and answer method of training being known as one of the renowned methods, still takes variant forms and stands out as better means of training.
Machine transits from one state to another. Being in one state, based on rules, we jump to next valid state. But what if we don’t know all the states yet? Is it possible to predict all the future states? How can we make sure that our machine has that all possible infinitely many states? There might be states which are not known yet and might be learnt through time.
The concept of abstract machine is no more only a mathematical fiction. If there is nothing new under the sun, then there is some unexplored area which might get the right principles to build a better machine. The basis of learning needs an exploration yet. The time variant and invariant rules demand a control machine which throws light on right time of applicability.
Machines have been trained like a child. It has got an expert system teacher. The machine having a clear state of mind is educated with knowledge and experience. The teacher gives rewards and as well as punishments. Machine being an obedient student believes that teacher is always right. Once it is trained it is then able to use its knowledge to make better decisions.
Machine however can never be like humans. Machine has its own limitations. Also, like said, “If each man had a definite set of rules of conduct by which he regulated his life, he would be no better than a machine. But there are no such rules, so men cannot be machines”. There are also other questions like, Do machines need emotions? What if humans lose them with time?
Since there is probably a very large number of satisfactory solutions the random method seems to be better than the systematic.
Alan Turing, Computing Machinery and Intelligence. Mind 49: 433-460.