May 09, 2018
Roger Melko, the Associate Professor at the University of Waterloo in the Department of Physics and Astronomy, has recently conducted a fascinating public lecture at the Perimeter Institute on the topic of ‘Artificial Intelligence and the Complexity Frontier’.
The lecture was to further develop aspects on how quantum mechanics added with Artificial Intelligence, can eventually create the complex frontier that could be impossible to breach even for the smartest humans. It also explored varied fields on how AI, with growing power of ‘Deep Learning’, can eventually reach a point where it can work on unsolved quantum physics mysteries, as good as, or better than, humans.
The beginning of Cryptography Decoding
In the beginning, Roger pointed out how in World War II, the famed cryptography machine ‘ENIGMA’ had already showcased technology’s aptitude in the field of intellect, and logical reasoning in order to surpass that of trained and really smart humans. The transistors on the cryptography machine were set in different ways every day to create coded messages that were impossible for any human to decode without prior understanding of the code book that was used to configure its settings.
In the end, though Alan Turing managed to break the code, he couldn’t do it without the help of another machine which he had built for that very purpose. He made the machine to calculate 10^20 possibilities in which the Enigma could be set for a day, and then cracked the code when the processed output made sense in plain English.
Emergence of Quantum Computing
It is comprehensible, that given enough circuits the sheer power of machines in quickly solving complex calculations, can create really useful applications. However, the sheer size of circuits was the most daunting issue to leverage the power of machines in the earliest ages of their abilities’ discovery. Then came Moore’s Law (The very precept that led the evolution of Apple and personal computers) and the observation that the number of transistors on an integrated circuit doubles every two years, entailing that the transistors would become smaller to be accommodated on a dense circuit.
This observation eventually led companies to evolve computing machines to the point where now a system as complex as smartphones, is now able to be fitted in your pocket, which otherwise would have acquired an entire city to function, not to mention the massive energy consumptions and radiations.
The same observation is now leading towards the age of quantum computing, that will use superposition of fundamental particles and their entanglement to relay information. This breakthrough, though currently in its infancy, is revolutionising our ideas when it comes to the processing of data, which thanks to the multiple factors has grown to become Big Data. It is noteworthy that AI is already being used to create miraculous predictions using the huge amount of data, by deciphering complex relationships between different variables.
Noting these observations, Roger pointed out how powerful artificial intelligence can actually become when its processing powers find expressions through quantum computing. Will machines then become able to solve complex problems that, until now, no human has managed to?
The Complexity Frontier
At this stage, where AI has enough of Big Data and processing power of quantum computing, the future becomes unpredictable for both humans and machines, creating the so-called Complexity Frontier. This future, according to Roger has a wide spectrum of possibilities. On one end, it will solve all the problems related to the ecology, politics, and physics but on the other, there is the pure dystopia. Furthermore, even if this advanced stage of AI does solve the profound mysteries, will humans be able to understand the solutions. Already the growing ubiquity of artificial intelligence for the sake of better ‘personalisation’ is creating discomfort among the public, raising privacy issues and consequently bigger problems like targeted psychological manipulation. It is hard to tell if we will ever become ready for the AI that is all-pervading on nano-scale, with intelligence far beyond any human’s. As Roger has brilliantly put it-
“We’ve taken what I thought the frontier, the sort of complexity frontier of quantum mechanics and we pushed it to some other level that we don’t even fully understand yet. So that’s my vision of the complexity frontier.”