Saturday, 20 September 2014

Is Big Data driving my car ?



This is happening today
Two days ago, while I was driving to go to a meeting, my car decided by itself to slow down! I did not ask for anything but apparently, my car “though” I was driving to fast!

On cruise control position, I don’t have to worry about my speed ! Not having any manoeuvre to do behind my wheel, I fall asleep but before my eyes close, the car would stop by itself on the roadside. And if I take a drink to wake up and go back driving, the car would start only after sniffing at my breath!

But, even if I am totally engaged and convinced for better living with Big Data, I feel somewhere uncomfortable.

Starting with dinosaur’s computers
Let us come back at the beginning of the movie, around year 1965 when I was a kid. 



Driving a car was quite simple: speed, breaks, gear, turn to the left, turn to the right…All this on mechanical commands and the driver deciding what action to do. Great ! Simple, straithforward. As a driver, from where did I get the information ? From the green light, from the policeman at the crossroad or from the signpost on the roadside.

Usage, responsibility and rules were clearly definned. No confusion



Shift of new responsibilities
Then, if we fly to year 2000, we found cars integrated with hundred of embedded systems taking control of isolated functions such as combustion, breaking, temperature…

For example, if I decided to break, my automatic systems would modulate my command for better effect. Ok, I can understand that an intelligent breaking systems, full of vitamins such as embedded systems and rocket scientist programmes could be smarter than my bear foots in sandals (or espadrilles, depends of the date) on a wet road.

But, responsibility between the driver and the car maker started to be confused. Still, the confusion was little because the embedded systems are closed systems having full control of the environment and the input/output data.

Now, who’s responsibility ?
Now, with Big Data, it is an open world, no more control on input/output of data. At year 2014, the car without driver is there, in front of you ! But, to my understanding, Big Data is more than that : to profit from the flow of data, if it is in an open exchange between the different sources of data.



Open world, losing control… frightening to loose control, don’t you think so?

Security for IT systems embedded in your car seems to be weak I can read in the morning newspaper. Virus on my breaking systems !! Gee ! Never think of it !

So Big Data is opening my car to let others drive it instead of me ! This is not a comfortable situation, you will agree with me!

As a user, my concern about Big Data is not to be afraid of loosing control of my vehicle (driving it, maintaining it, replacing it…). My concern is what service I what should I pay for it !! Is Big Data serving my needs or serving others interests…

Shall I give the responsibility (and the pleasure) to a Machine Learning ?
Big Data in coming with a lot of data. From my understanding, traditional computing techniques are not able to deal with Big Data and this open the doors to Machine Learning

And the question is : would you let a Machine Learning algorithm take control of your driving? Stanley Kubrick brouth the question in “2001 Space Odyssey” with HAL computer. And the answer was: pull off the plug. But today, it is impossible to pull off all the plugs of Internet, the back bone of Big Data. We are surrounded by an ever running Machine Learning. For who’s benefits this Machine Learning is working for (hard working, truly !) ?

Should I leave my driving skills to a Machine Learning ? Should I leave my pleasure of driving to a Machine Learning ?



So if you want to continue driving by yourself and enjoying by yourself doing things (like at the time of the roman empire) I only see one solution : get control of the Machine Learning.

But how to get control of a Machine Learning ? Let’s wait for the next post. I shall have a cup of tea in between

Cheers
Robert


No comments:

Post a Comment