AI powers machines to win over mankind at Poker. Machine Learns to Introspect.

0
693
AI. artificial intelligence, artificial, intelligence. machine, learning, machine learning, ML, carnegie, mellon, computer, intelligence, algorithm, poker, tournament, GO, chinese , game, chess, Deep, Mind, google, illusion, maya, indian, philosophy, oriental,

AI. artificial intelligence, artificial, intelligence. machine, learning, machine learning, ML, carnegie, mellon, computer, intelligence, algorithm, poker, tournament, GO, chinese , game, chess, Deep, Mind, google, illusion, maya, indian, philosophy, oriental,

AI wins Poker tournament

AI (Artificial Intelligence) is going places. The latest instance is a software, Libratus beating professional Poker players in a tournament. The program created by researchers  from Carnegie Mellon University beat some of the world’s best Poker players.  

* Libratus’ victory was made possible by the Pittsburgh Supercomputing Center’s Bridges computer, on which the AI computed its strategy before and during the event, and by Rivers Casino, which hosted the event.

Throughout the competition, Libratus recruited the raw power of approximately 600 of Bridges’ 846 compute nodes. Bridges total speed is 1.35 petaflops, that is about 7,250 times faster than a high-end laptop. Its memory is 274 Terabytes, about 17,500 as much as you’d get in the above laptop. This computing power gave Libratus the ability to play four of the best Texas Hold’em players in the world at once and beat them.*

Machine Learns to Bluff 

One of the key elements in card games like Poker is the need to bluff to unsettle opponents or lead them the wrong way. If the machine cannot learn to bluff and bluff well, it decreases its own chances of winning. This is a critical difference between traditional mind games like Chess or the ancient Chinese GO. 

Games like Poker entail continuous learning which could be person dependent as well, since the way one player bluffs would be much different from another’s, with the same set of cards. It is not just about playing your turn basis the set of cards you have or the cards played by the other player. Programs playing games like Chess and GO don’t have this additional nuance/ dimension to consider. The software just has to calculate all the possible moves and take a call basis the best possible outcome from multiple ones. 

Machine Learns to Introspect 

Programs developed to play Poker till now have focussed on how the opponents played and worked to exploit their weaknesses. Libratus has turned this ‘occidental’ thinking on its head. It has learnt to play better by introspection. 

Introspection, the latest ally of AI

As the end of each day’s play, Libratus would list its own weak moves of the day. Of these weak moves, it would take the three weakest ones which enabled opponents to win. The algorithm would then tweak itself to avoid the same mistakes. Over a period of twenty days, Libratus learnt enough and tweaked itself enough to beat four of the world’s best Poker players and win the tournament.

The researchers opine that Libratus’s abilities (of learning to bluff and introspect) could have far reaching applications. A simple example given is that of our smartphones negotiating the best price for a new car in the future.  

Cool Chilli’s Comments 

Be the Change

Introspection and Meditation has been a long standing tool for self development in ancient Indian/Oriental Philosophy. The belief was/is that man and woman are part of the cosmic consciousness. Meditation and Introspection takes us closer to that sublime state of Godliness. The more we see within, understand the changes needed and try to bring about the changes the better we get. Be the Change That You Wish To See, as Mahatma Gandhi beautifully put it.

Practical Benefits of Meditation

Cool Chilli‘s friend, Ranganathan follows the ancient Brahminical tradition of Sandhya Vandhanam. ( Prayers at/to Twilight literally) to the T.  Performed thrice a day, these prayer rituals sublimate the mind, facilitate clear thoughts and bolster memory. An important part of these prayers is Meditating on the Gayatri Mantra.  Meditation was encouraged for adolescents of all castes.

Ranga does not need to store contact numbers on his mobile.  More than 500 numbers of his family, friends and business associates are clear in his mind,  as is his conviction that Meditation ‘does the trick for him’.

We see that the newer generation kids finds praying for 5 minutes a challenge. Meditating for 15-20 minutes is remote. These kids otherwise smart, capable of coding even apps, may reach for a calculator, for the arithmetic that their grandparents did mentally. Ranga’s experience rings true!  

Little wonder that copious amounts of knowledge and wisdom were handed down in India through generations of sages by oral tradition only. ‘Why bother to write it down?’. It is all in the mind! Naysayers may have a point that without written documentation, knowledge sharing cannot be validated for quality etc.  

In Ending          

In a world where even computers and machines are getting more EQ and IQ (learning to bluff well is not just IQ), SQ (Spiritual Quotient) is still majorly unknown.

But for one with enough SQ, winning or losing may mean the same, since He or She just goes with the flow. For these evolved souls, the world with all its events and outcomes is just an illusion! Sab Maya Hai!

Related Reading and Acknowledgements

*Carnegie Mellon University note