Do You Revisit Your Old Ideas?

…I did from 27 Years Ago

Robert Trajkovski
3 min readMay 5, 2021
Photo by Possessed Photography on Unsplash

After finishing my Masters in Control Systems and Robotics I decided that I wanted to pursue my doctorate. I pursued the PhD degree with focus on AI from 94 till 99. Life got in the way and I never finished it.

However…

What was I trying to do was very unique at that time. And I believe today. I came up with the idea that data was becoming prevalent but access to experts was going to be difficult to obtain.

With this in mind I worked on combining ideas from Data Mining, Machine Learning, Artificial Neural Networks, Neural Fuzzy Control area. The idea was to go from data to rules that are extracted.

At that time I ran an experiment on a piece of data that had 30 variable of which 2 were outputs and had 3000 rows of those 30 variables. Data was from a blast furnace heat in a steel mill. That data was to be transferred to another operator at the Basic Oxygen Furnace to make a decision as to what to add to the iron to turn it into steel with certain carbon and hardness levels.

After working on this problem for a while, I finally solved it and produced rules that an operator can reference as recommendations or an automated system that can make those decisions in real time.

After I stopped pursing the PhD. degree I never did theoretical research again. I just moved on. BUT I often wonder if I should have pursued it and where I would be today. It is not something I dwell on BUT just wonder.

Recently I bought a book called The Master Algorithm. An excellent book that teaches about machine learning and the different schools of thought within it.

While reading it I kept recalling that most of the ideas that those schools explored were also explored by me between 95–99. There was only one new idea that did not exist at that time or I was not aware of it at that time.

I patted myself on the back BUT then I realized that I did not complete my work. The field still does not have a unifying theory for learning from data. With different types of data different school algorithms would be more relevant.

So there is work to do.

Last night I was reading a book on game theory and competition. I immediately saw how some of the ideas from game theory with fuzzy logic could be used to improve genetic programming. It was hard to sleep with these ideas floating in my mind.

Am I telling this story to brag how smart I am? Nope.

The intent of the story is to realize that sometimes our ideas are ahead of time. Often many years ahead of time. We stop their pursuit because life gets in the way or something else looks sexier. This could be a mistake.

I do believe that now I can see how a additional areas of study could be added to my old thinking process to hopefully improve the data to information process. Should I pursue it is the big question?

Because there are rarely re-dos in life, maybe our motto should be FOCUS (from JLD on Entrepreneur on Fire podcast)

Follow

One

Course

Until

Success

Following multiple courses and ideas drains our time and reduces a chance for success.

My four cents…

--

--

Robert Trajkovski

I have led people and projects in Steel/ Power, Refining, Chemicals, Industrial Gasses, Software, Consulting and Academia. I have instructed 73+ courses.