Little nightly thoughts
by Business Exploration®
The 2nd Scientific Revolution
We are not in the 4th Industrial Revolution.
We are in the 2nd Scientific Revolution.
Dear Fellow Innovator,
In Today I was at the Dubai Data Science Meetup in Deira City Center, an amazing group of "nerds" in love with statistics and business.
I discovered new things, like:
· how to map processes without mapping them
( thanks to Safdar Hussain who showed a machine learning algorithm capable to draft by itself a process map, starting from the logs of company transactional data)
· how to use a self-learning algorithm to pilot an helicopter
(thanks to Vitalii Duk who showed that it is not only possible but even simple for a machine to learn to control a complex system if proper rewards are instructed )
But the most important thing is that I finally understood what Data Science is.
Let's start with debunking a myth:
We are not in the 4th Industrial Revolution era.
Moreover: The 4th Industrial Revolution is not about Connected Automation, IoT, Internet, M2M, Data Analytics, etc etc.
But it's true: something happened in the last few years. It's
The 2nd Scientific Revolution.
We are in a new era, and probably we are not fully grasping it.
What sets a part these few years from the last 4 Centuries?
We changed the way we solve problems.
In other words we changed the way we innovate and this means a further step in our (our?) evolutionary journey.
The last 400 years we benefited from the discovery of a powerful methodology for inventive problem solving:
the Scientific Methodology.
Galileo Galilei and Sir Francis Bacon gave us the tool to take better decisions:
Find a Model of Reality and apply its General Laws to solve the actual problem.
It worked very fine.
It leads us step by step from governing mechanics, thermodynamics, to chemistry, electromagnetism, and nuclear forces.
Creating a General Model and applying it to make actual decisions has become the way we live and have lived till ...
the advent of the microprocessor and its amazing calculation power.
Here the path of innovation diverge.
Organizations like the Financial Institutions start to use this amazing power to make decisions based on the chances to optimize the results.
The game changed.
The way we innovate (solve decisions problems) was no more based on the identification of a general model, to identify general laws, but on the:
heuristic approach in finding the path with the higher chances of success.
We started deciding what to do, trying millions of times untill we have enough confidence that the solution A gives better results than solution B.
After the Banks, we started to use the heuristic approach to support decisions based on how many chances there are that a piece of information deserves us, if it has deserved other persons in the past.
We called it "Google Algorithm".
Then we started to use it to support our buying decisions, with the use of clustering techniques to drive show what others have bought in the past.
We called it "Amazon reviews".
Today we are moving forward: the heuristic approach is used to make decisions in our behalf.
We design software that fails so many times that it can find the right way to maximize a result, totally by itself.
(The cost of experimentation has gone to nothing, and we took advantage failing milions of time.)
This is pervasive.
Every time we use a clustering technique, a classification model, that we bet on an optimal choice based on its probability to be a better one given past trials,
we use a new paradigm to make our decision:
And we are starting doing it every day: when we choose a book on Amazon, when we optimize our investments, when a Company decides pricing policies or when a car anticipates the mistakes of its driver.
Those Countries that have been able to embrace the new Scientific Methodology, have been capable of massive productivity gains: USA, UK, France, Germany.
The other, like Italy, are lagging behind.