The recording of this inaugural session of PSYCHE along with the opening lecture by Stelarc can be seen at Split Body, Synthetic Self: Excess and Emptiness | Lecture - YouTube
Good evening all and thank you Jahnavi, Madhu, Yamuna and the rest of the Science Gallery team for this excellent new exhibition and also inviting me to make the opening remarks.
I am trained in computational mathematics which I have also been applying to the life and health sciences over the last couple of decades. When I heard about the plans for Psyche which I understand as the totality of elements forming the mind. My first thoughts went to recent conversations with the investor community in biotechnology who were really excited about the mainstreaming of psychedelics as pharmacology for mental health issues.
So when Jahnavi asked me to give these remarks I thought I could be a “new age neuronaut” and become the Timothy Leary of this millennium. On hearing my plan, a friend even advised “just pop a pill and see how it goes!”
However, my intellectual conscience steered me towards my own training in computing and so instead of psychedelics I will briefly examine the question, ”IS MY PSYCHE COMPUTABLE?”
There is a good reason to ask this question which stems from the tradition of “Computational Positivsm” that defined Classical Indian Science. This tradition is well researched and documented by my intellectual mentor and hero, the late Professor Roddam Narasimha. Two decades back, when I served on the faculty at the National Institute of Advanced Studies that he headed, I am sure he would have asked me to speak on this topic if it had arisen in conversation. So these remarks are in some way my eulogy for Professor Narasimha.
In order to address the question “IS MY PSYCHE COMPUTABLE” I went back to the master – Alan Turing and re-examined his treatise published in MIND in 1950 in which he raised the question “Can Machines Think?”.
Most of us remember this work of Turing as the paper that introduced the idea of “The Imitation Game” played by a Human, a Computer and an Interrogator. The objective of the Human is to get the Interrogator to choose correctly who answered as a human and who as a machine while the objective of the Computer is to get the opposite outcome. There is a lot more detail but that is more or less the gist of it.
This came to be known as the “Turing Test” and many claims and counter claims have been made about computers at IBM and Deep Mind passing the test in the domains of Chess, Jeopardy and most recently with Go with flying colours.
Well, as I re-examined the paper, I found that Turing had delved a lot deeper than these games and his insights on “Can computers think?” basically gave me some clues on my question, “Is my psyche computable?”.
Turing laid out a list of possible objections that may be raised on his claim that machines can win the imitation game once the capacity of machines reached the ability to process 10^9 bits of information (approx. size of encoding Encyclopedia Brittanica). Incidentally, some 60 years later, scientists at TJ Watson at IBM did something similar – they trained “Watson” to read and digest the contents of Wikipedia before challenging the world champions in a game of Jeopardy.
The list of objections included
- The Theological Objection (No immortal Soul from God for a Machine)
- The Heads in the Sand Objection (Too dreadful - Hope this never happens)
- The mathematical Objection (Godel’s Theorem and Undecidability)
- The argument from Consciousness (write a sonnet or compose a concerto with feelings) – Francis Crick and Roger Penrose had a lot to say about this later.
- Arguments from disabilities of Computers (enjoy strawberries and cream)
- Lady Lovelace Objection (Analytic Engines can only do what it is instructed to)
- Analog vs Digital Nervous System Objection (differential analyser)
- Informality of Behaviour Argument (following rigid set of rules makes a man no better than a machine)
- The argument from Extra Sensory Perception (telepathy, clairvoyance, precognition and psycho-kineses. Turing notes that if clairvoyance etc are real, the Turing Test will need to be tightened up. For example, competitors have to be placed in telepathy free rooms!
Turing ends his extraordinary treatise (often considered the real foundation stone of Artificial Intelligence) with an exceptional idea. The idea is to create a child computer and then make it learn. And Turing goes on to outline how machine learning may be facilitated. Remember this paper was published in 1950!
Well, I shall leave it there and ask you to go read the original[1] and enjoy the rest of the program on PSYCHE that Science Gallery Bengaluru has curated brilliantly for you beginning with the opening lecture by Stelarc.
Thank you.
[1] Turing, A. M. (1950). Computing Machinery and Intelligence. Mind, 59(236), 433–460.