Algorithms and AI by Jim Moore

Algorithms and AI Artificial Intelligence

Jim Moore’s Air Force Technology Experience

Jim and Ying Moore have been our neighbors here in PebbleCreek for many years. When we first met a few years ago, we learned that we all lived a mile apart in North Bend, Washington. It’s a small world after all.

We quickly became friends while sharing our life's journeys over lunch, dinner or just sitting around gabbing. I shared my writing of this “book” and asked Jim if he would be interested in sharing his knowledge and experiences. He agreed and he will continue to write until he or his family and friends decide not to. Thanks for sharing!

My Experience with Algorithms by: Jim Moore

Algorithm is a fancy word for what my mom would have recognized as a “recipe”

It specifies the inputs (ingredients), the procedure for converting them (baking instructions) and the outputs (cake). However, most of what goes on in computers is not visible since processing is done in terms of individual 1s and 0s. This lack of visibility is what allowed ‘programmers’ to attain cult status in the 50’s and 60’s. Whether they worked for NASA or your local bank, the mystique was enhanced by the wearing of plastic pocket protectors full of pencils and pens. Pocket protectors were the first badge of “geekdom”.

And so, as I ride off into the sunset of retirement from the computer industry, Artificial Intelligence (AI) and in particular machine learning is the next great frontier. Algorithms are to AI as DNA is to biology. AI gives computing systems the ability to learn and respond independently without the need for human intervention, changing their algorithms on their own and assessing the results. Then changing again. We will perhaps be able to see what the machine does, but not how it decided to do it. Just last year there were ten ‘breakthrough’ algorithms introduced to computing. As the robots get more advanced, they may eventually command us to “Step away from the machine, human!”

As a fledgling “system programmer” I worked part-time in the Ohio State University Data Center with the biggest computers of the day. In the middle of the night and on weekends, we would communicate with the inner workings of the IBM computers in hexadecimal codes. I felt a lot like the sorcerer’s apprentice in Disney’s Fantasia cartoon.

Then,

Always working in the middle of the night, I grew comfortable communicating with electronic beings. It turns out that the very biggest computers were quite similar to the very smallest in many ways. When General Motors began to design computer electronic control modules for cars, I was hired and moved to Detroit.

The word algorithm itself is derived from the 9th century mathematician, Algoritmi. A partial formalization of what would become the modern concept of algorithm began with attempts to solve a decision problem. The concept of algorithms has existed for centuries. Greek mathematicians used algorithms for finding prime numbers, and for finding the greatest common divisor of two numbers. Precision, especially in regard to inputs, was key to success. “Garbage In, Garbage Out”. Thanks to NASA the term “glitch” became widely recognized. What was not so widely recognized, though, was that glitches were almost always caused by human error. Just like in baking, if your mom mistakenly used avocados rather than apples the pie was not at all what we expected. If the oven was set at 500 degrees rather than 350, it was also a disaster.

Algorithms were tireless but dumb. They did EXACTLY (and only) what they had been told to do. And the computer (or the car) always got blamed for it - not the original author engineer/programmer. Life was good!

Then,

By the late-1980s, computers and electronic devices began to get smarter. They corrected your spelling and began to guess what you were going to type as the next word in the sentence. You even got to choose which word. Suddenly, the inputs didn’t have to be as precise as in the old days. Anyone could text or even tell a computer (including iPhones) what to do! Everybody could “program” something. Today, you can talk to your TV’s remote. Pocket protectors went into well-deserved retirement. Everyone dragged around chargers and cords instead.

Then,

Programmers weren’t so special anymore, so I became enamored with “data”. There was just so much of it! Computers churned out more and more every second 24/7. I became a ‘data scientist’ and went to work for IBM. I bought blue suits, white button-down shirts and red ties. As we learned more about managing data, our clients ranged from oil exploration companies to the International Olympic Committee. Everyone was convinced that there was “gold in them thar (data) mountains”. And they were right. ‘Data mining’ became a thing.

Enabled by further advances in the underlying technologies, most of the rigid restrictions on input data have now been removed. Something called “big data” appeared. (Obviously, not much thought went into the name.) But it means that large and complex volumes of all kinds of data can serve as input to more complex analysis and prediction algorithms. The result has advanced everything from cancer research to Amazon to my own home life. My Roomba vacuum finds its way around the house – even if I rearrange all the furniture. When the battery gets low, it finds its way back to the charging station and plugs itself into the receptacle. Hopefully, soon it will learn not to come in the bedroom when I am sleeping or the living room when the football game is on. And Wahoo! Because of data and some clever algorithms, my car can parallel park itself!

Even a different kind of refrigerator is almost upon us. It will be able to identify and track each item inside and their expiration dates. Based on the contents of the refrigerator, we need only to give it the number of people to be served plus the start time for dinner and it can suggest all the possible recipes to make a full meal. My Mom seemed to spend half her life planning meals. I can see her smiling now.

Then,

Up to this point in the chronology, computers and technology have required humans to initiate the algorithms by doing, touching, or saying something.

“Alexa, …”.

But what if you could just mentally WILL them to do something? An Australian company recently introduced a headset that serves as a brain-computer interface. Wear the headset, think about what you want to do, and you can control games on Windows, OS X and Linux operating systems. Or change channels on your TV. Or turn off the lights.

“What was I thinking?” takes on a whole new meaning.

But nothing is completely good. “What was HE thinking?” opens up a whole new market for spouses. Imagine the chaos introduced if they could read your mind in real time.

Then,

Up to this point in our story, one or more algorithms (really just recipes, remember?) are behind each of the advances I have cited. Once they were written, encoded, and loaded they stayed the same until changed by again by a human being.

A common theme in science fiction has for decades been the ‘rise of the machines’. In this scenario humans lose control of the algorithms, and the machines take on a life of their own. We may have just passed that tipping point. Cloud computing is a term for both complex global system designs that we “… needn’t understand in depth” and massive arrays of actual technologies that only a very few can observe in their entirety. And really no one completely controls.

So, I began to aspire to be a high priest of Cloud Architecture. It was in response to a last-ditch effort to put some management boundaries around the ever-increasing complexities of networked technologies. IBM’s marketing pitch at the time was, “IBM Cloud / The Cloud that Manages It All”.

Well, maybe. IBM’s Watson does win at World Chess Cups and Jeopardy. But the possible permutations of components in the cloud are totally mind-boggling and perhaps beyond our intellectual capability to control. Things were so much simpler when we could only wear blue suits, white shirts and red ties to work.

What happens when these systems go out of control? Here is an example - High Frequency Stock Trading has caused numerous ‘flash crashes’ on Wall Street in the last few years as dueling algorithms competed at lightning speeds. Human traders had no hand in these billion-dollar exchanges of value, but there were clear winners and losers. Probably not you or I though.

Page updated

Google Sites

Report abuse