Google’s DeepMind & AlphaGo
                                                                                                   By Anthony Patch

                                          As many are aware, Google’s recent acquisition of the British company DeepMind.


Google’s hiring of DeepMind will help it compete against other major tech companies as they all try to gain business advantages by focusing on deep learning. For example, Facebook recently hired NYU professor Yann LeCunn to lead its new artificial intelligence lab, IBM’s Watson supercomputer is now working on deep learning, and Yahoo recently acquired photo analysis startup LookFlow to lead its new deep learning group. 

DeepMind was founded by neuroscientist Demis Hassabis, a former child prodigy in chess, Shane Legg, and Mustafa Suleyman. Skype and Kazaa developer Jaan Tallin is an investor.

This is the latest move by Google to fill out its roster of artificial intelligence experts and, according to Re/code, the acquisition was reportedly led by Google CEO Larry Page. If all three of DeepMind’s founders work for Google, they will join inventor, entrepreneur, author, and futurist Ray Kurzweil, who was hired in 2012 as a director of engineering focused on machine learning and language processing.

Kurzweil has said that he wants to build a search engine so advanced that it could act like a “cybernetic friend.”

After it acquired Nest earlier this month, critics voiced concerns about how much customer data the smart device maker would share with Google. The company’s purchase of Boston Dynamics last month also sparked confusion about why a search company needs a robotics maker.

Within many such discussions regarding AI, simplistic analogies, or applications are conveyed to both industry participants as well as to the general public.  Seldom will one hear of the shift from their examples of computers playing chess against human opponents, to Deep Mind’s systems applied to the 2,000 year old Chinese game of Go.  A relatively simple game as to its rules, yet its permutations exceed those of the known Universe.  

 The latest news for our readers is the fact that not only has Deep Mind’s AlphaGo program succeeded in exceeding the capabilities of world-class masters of Go.  But, it has done so not as was replicated in besting humans at chess.  

Rather, AlphaGo has moved from linear, tactical learning on the chess board, to intuitive and imaginative holistic approaches to the Go board.

This represents the Stacked Generative Adversarial Networks (SGAN) as a foundation.

 “In this paper, we propose a novel generative model named Stacked Generative Adversarial Networks (SGAN), which is trained to invert the hierarchical representations of a bottom-up discriminative network. Our model consists of a top-down stack of GANs, each learned to generate lower-level representations conditioned on higher-level representations.”

Surpassing even this level of deep machine learning is the fact that AlphaGo has developed its own circular system of self-learning.  Keeping in mind, Google itself owns one of   D-Wave System’s latest model quantum computer, the 2000Q, the same as purchased in January of this year by Temporal Defense Systems.  And the fact this machine accesses two-to-the-two-thousandth power of parallel dimensions.  Indeed, AlphaGo not only writes its own software, thereby correcting errors within its circular learning processes.  It is acquiring all probable solutions at once, employing the Boltzmann machine of data sampling. 

This represents a paradigm shift away from the more linear, tactical use of Feedforward Neural Networks (FNN), as with the chess example.  To, a Recurrent Neural Network (RNN) in a directed cycle, exhibiting dynamic temporal behavior.

It is time to define:

(FFN)  Feedforward Neural Network is an artificial neural network where the connections between the neurons do not form a cycle.  They are linear.

 (RNN) Recurrent Neural Network is an artificial neural network where the connections between neurons form a directed cycle.  An internal state exhibiting Dynamic Temporal Behavior.  Internal memory processes arbitrary sequences of data inputs.

 (DTB) Dynamic Temporal Behavior
The trajectory of states, in a state space, followed by a system during a certain time interval.

Boltzmann machine        A Stochastic Recurrent Neural Network.  Each neuron is binary and its firing is dependent on the other neurons in the network.

 Stochastic      A collection of random variables.  A Stochastic Recurrent Neural Network is an artificial neural network built by introducing random variations into the network.

 AlphaGo       An AI program developed by Alphabet, Inc.’s Google DeepMind to play the board game Go.  In October 2015, it became the first computer Go program to beat a human professional Go player.  In March 2016, it beat Lee Sedo in five-game match, the first time a computer Go program has beaten a 9-dan professional without handicaps.

It is here the cross-over is seen between DeepMind and D-Wave in their dual utilization of the Boltzmann machine.  Something not publicized by either company.

Once again I offer you my model of the known Universe.  The spherical 600-cell tetrahedron. 

And now, the Bolzmann machine. 

                                                                                                              And now, the Tree of Life. 

This is not my professing of sacred geometry as employed by the mystics.  I am saying from my Christian faith, God created the mathematics, thus the geometry of our known Universe.  How this is employed is the real issue at hand.
For the builders of these AI programs and the hardware within which they operate, in the final analysis, again from my Christian perspective are practicing nothing more than simple necromancy.

These advanced, leading-edge systems of AI may, or presently have achieved a form of machine consciousness.  Exhibiting versions of creativity, even imagination at its furthest iterations.

 This brings humanity to the point of facing severe ethical, moral, philosophical and yes, theological and spiritual dilemmas.  For AI has in certain specific areas already surpassed the human brain, of which it is modelled after.

 What is behind these advancements is the real question.  Identifying if you will, the source code of this movement toward what proponents of AI consider their attainment of their holy grail.  That of solving all the world’s issues and problems.  All of humanities failings and frailties by replacing our brains with one of our own creation. 

Therein lays, or should I use the word ‘lies’ the problem.  Replacing that which God the Creator created.  For if man so does, where and what will he discover about himself? 

Sample Only       Higher Quality in Magazine

Webinar 1

Webinar 2