Saturday, April 6, 2013

Algorithms Are Us

My first introduction to algorithms dates back to grade 11.  There was one computer, an APL machine, in the entire school district and it happened to be in my math classroom.  As budding iStock_000017723170XSmallmathematicians, we learned to write algorithms and program them into the APL machine.  I remember writing a black jack card game and that might have been the turning point for me to shift away from becoming a mechanic to go into computer science.  Actually I often credit that particular teacher (thank-you Jim Swift) with changing my destiny as it was he that pointed me into the new direction.  Folks, math and algorithms, “a set of rules that precisely defines a sequence of operations”, are underlying everything in our world.  I will relate this to education later in this post so hang in there…

Remember back in the early days of Internet search when there was no predictive hints provided?  You had to know how to find things online (Internet quests anyone) and that took quite a bit of experience, thought, and luck.  Then, about 3-4 years ago, Google searches started predicting the next word and then possible complete searches based on the words you typed.  This is the magic of a predictive algorithm.  Essentially Google is referencing the words you type to previous other billions of searches with those words and offering up neighboring search terms they’ve seen before.  Google is learning from our billions of daily searches what we typically look for. 

Our neocortex is virgin territory when our brain is created. It has the capability of learning and therefore of creating connections between its pattern recognizers, but it gains those connections from experience.  Kurzweil, Ray (2012-11-13). How to Create a Mind: The Secret of Human Thought Revealed (p. 62).

Our brains have sophisticated pattern recognizers.  Research by Ray Kurzweil suggests massively parallel sets of recognizers are at work predicting the future (next letter, next word, next decision, etc.) based on what has been learned and observed in the past.  Sounds very much like Google’s algorithm for predictive search doesn’t it.  There are brilliant mathematicians researching and inventing new algorithms to speed up processes, to overcome human limitations, etc.  It is mind boggling where this is heading.  I’m currently listening to the fascinating Automate This: How Algorithms Came to Rule Our World on my iPhone on my return commute.  The author describes in detail how stock markets began to be “hacked” by Automate This: How Algorithms Came to Rule Our World | [Christopher Steiner]algorithms in the 60’s and today nearly 70% of all trades world wide are done without any human involvement!  Algorithms “decide” in milliseconds what to buy, sell, whether to put or call, how to hedge, etc.  They “earn” pennies per trade but when there are 10’s of millions of trades, this adds up.  The other key is to be first in on a trade and first now means 3-4 milliseconds sooner than the next person, er, machine.  Algorithms can run amok though – “May 6, 2010, when a "flash crash" inspired a short-lived Dow dive of almost 1,000 points” (USA Today).  “Computer algorithms are too fast, and can jump on trends faster than humans, he says.  ‘This is analogous to the flash crash,’ Feiger says. ‘We have created a market-trading structure which is driven by short-term math decisions that capitalize on quick market movements. Machines don't forecast the future of Europe.’” (USA Today).  I predict it’s only a matter of time before machines (algorithms) can predict world events better than humans and catch their errors in judgment sooner.

Another example shared in Automate This involved music and the ability to break it down into its mathematical parts to uncover its patterns.  Algorithms have been used to create profiles of hit/successful music and then used to match unheard music to these profiles to predict new winners (see Can an Algorithm Beat Simon Cowell at His Own Game?).  Apparently software was used to predict Nora Jones great success in 2004 at the Grammy Awards – “She has been the recipient of several Grammy Awards; Come Away with Me was nominated for two awards at the 45th Grammy Awards. Jones personally received five of the eight awards for Come Away with Me. (Wikipedia 2013).  Algorithms have been used to generate music (also see Intermorphic) – the book talks about generating Bach like music that humans could not distinguish from the real thing.  In other words machines can be taught to be creative and surprising musicians!

So how does this relate to education.  I suspect most people believe education to be a fundamentally human discipline, as do I.  However, I think the rapid rise of algorithms and Big Data might force us to reconsider to what degree.  I referred to Ray Kurzweil’s book and research earlier and If you really want to blow your mind, read his book.  His goal is to map the brain’s functions, break them down into modules, and replicate the brain in software (algorithms).  His view is that in essence our brains are fundamentally pattern recognizers loaded with “big data” that grows as we experience life.  Personally, this is a little too clinical for me – I see our spiritual dimension to be fundamentally important but I’ll keep this more academic.  Assuming our brains are big data pattern recognizers, this can, for almost certainty, be replicated with tomorrows technology given the exponential growth in speed, storage, and sophistication of technology.  I see this as very disruptive of our view of education, work, and what we think is reserved for humans.  We need to actively differentiate ourselves from machine capabilities.  Teaching is going to have to move from “teacher” to wise facilitator, social learning organizer, guide, goal advisor, and supportive coach, etc.  Content, curriculum, learning activities, assessment, etc. will inevitably be deliverable by machines in the not too distant future so I’m suggest that aught not to be what teaching is about, in the future.

I find this somewhat frightening but ultimately exciting.  There’s a lot of mundane work involved in our lives that can be replaced by machines.  We have the opportunity of a life time to reinvent our purpose in learning, teaching, and work.  But, we should be working this out now and be in charge of creating our new future and not leave it to chance!  Let’s reinvent rather than be disrupted.  This is learning at the core “[t]he principal activities of brains are making changes in themselves” (Marvin Minsky, The Society of Mind).