If youâ€™re looking for proof of how Artificial Intelligence and Machine Learning technologies are seriously challengingâ€¦
And worryingly for humans(!)â€¦
Even now surpassing human intelligence, guile and ingenuity then look no further than Libratus.
Built by computer science experts at Carnegie Mellon university, Libratus made poker history in January 2017 by taking on world class professionals at No Limit Texas Holdâ€™Em and beating them over 20 straight days.
This Man vs Machine battle has been done many times before going back to when chess grandmaster Gary Kasparov played IBMs Deep Blue in 1996.
But this Libratus experiment was different.
Because Libratus used something called â€˜reinforcement learningâ€™ â€“ basically, learning from scratch and then using trial-and-error to get better and better.
In fact, Libratus learnt and improved before, during and after each dayâ€™s play. It literally had an unfair advantage over the four players it challenged.
And whatâ€™s moreâ€¦
And this is the really scary bitâ€¦
It knew how to bluff and when its opponents were bluffing!
As Wired said in their article of February 1stâ€¦
â€œItâ€™s a powerful and rather unsettling proposition â€“ a machine than can out-bluff a humanâ€.
I dread to think whatâ€™s next for Artificial Intelligence and Machine Learning!
Steve Clark is the founder of Winning Online an information publisher with a speciality in algorithmic betting. Steve can be followedÂ or contactedÂ at https://www.betgeeks.com/members/steve-winning-online/profile/Recommend0 recommendationsPublished in