“The question of whether a computer can think is no more interesting than the question of whether a submarine can swim.”
I and Pavan, a man with extensive experience in making trading algorithms, were debating – In a world filled with smartest traders and smartest algorithms, who can manipulate better to win. The answer came to me last weekend while reading Deep Thinking by Gary Kasparov.
Because algorithms learn from a limited history(with no creativity), they will make the same mistakes as men.
Since uncertainties happen all the time, unsupervised algorithms will be baffled when they see new conditions.
- Men make algorithms and this limits their ability to understand the world. Hence, either algorithms are always one step behind men or perform equivalent.
- Men can manipulate the algo because he knows beforehand how an algorithm will behave since he knows its working principles
A computer algorithm written by men to play chess beats world champions till date.
Stock market is similar but it’s a multi-agent trading system – which increases complexity of calculations even more – where computers triumph(if they have all the information – numerical and text)
In respect to creativity and ability to manipulate, AlphaGo made a move which was considered impractical by Go world champions. This move was bizarre but actually led to win. This proves the capability of machine learning algorithms to understand the rules, patterns and take unforeseen actions which will not make sense to us.
Computers are not as limited by historical data in thinking as we think of them. Algorithms like ICA/PCA and Neural networks can help them extract features which are not apparent to men. Our inability to predict their ability comes more from the limitations of our own imagination.
The book is going fantastic so far and interestingly its caption says – Where Machine Intelligence Ends and Human Creativity Begins. I am looking forward to how this changes my opinion on limitations of AI.
“I believe that at the end of the century the use of words and general educated opinion will have altered so much that one will be able to speak of machines thinking without expecting to be contradicted.”
1985 – Kasparov won by 32-0 against 32 computers simultaneously
1989 – Kasparov won by 2-0 against Deep Thought
1996 – Kasparov won by 4-2 against Deep Blue
1997 – Kasparov lost for the first time against a computer by 2½-3½. It was the comeback of IBM’s Deep Blue.
An AI evangelist and a multi-disciplinary engineer. Loves to read business and psychology during leisure time. Connect with him any time on LinkedIn for a quick chat on AI!