Logical Religion 4.0: Can we predict the future?
Consider that there are six billion people, each living on Earth, with a finite set of characteristics and each of those characteristics having a finite set of values. We know the outcome if a few of those characteristics meet in a known situation.
For example, in a gross simplification, if there were only two people in the world, Vinit and me, the situation being of us walking to our office in the morning. We both have a set of characteristics; a few of the relevant ones in this scenario, being:
- ST's Elevator preference: Very Strong Affinity
- VB's Elevator preference: Weak Affinity
- ST's Staircase preference: Very Strong Aversion
- VB's Staircase preference: Strong Affinity
Just with these 2 sets, we know what the outcome will be: we take the elevator.
Now, we know that we will take the elevator and we know the time it will take us to go up the elevator, so, we have not only predicted the immediate future but also made a base for the next prediction: of what will happen once we get out of the elevator and so on.
If, only Vinit and I existed in the entire world with only say 10 characteristics each, we could easily predict our entire lives, even with the current computer powers and our limited brain capacities.
The French scientist, Laplace had a similar idea but for predicting the state of the universe:
Laplace suggested that there should be a set of scientific laws that would allow us to predict everything that would happen in the universe, if only we knew the complete state of the universe at one time.
(Stephen Hawking, A brief history of time)
The problem with Laplace suggestion was, of course, that by Heisenberg uncertainty principle, we can not know both the momentum and current position of any particle. Thus, effectively ruling out the possibility of us knowing the state of the universe and thus predict an earthquake or a black hole forming, even if, we had a HUGE computer.
But we don't have that problem.
Humans have a finite set of characteristics whether the number is 1000, 5000 or 20,000. Humans themselves are finite in number; we don't know what the exact number is right now, but that doesn't mean that we can't find out. Each human interacts with a finite number of other humans, multiple times and with another finite number just once, over the course of her life. Some of the characteristics are weaker in some rather than in others.
For example, my friend J never had Indian food before and thus THAT characteristic of her was 'weak' or 'empty'. If she were to meet people who hate Indian food, her characteristic of Indian food would get filled with their's. But since she met us, our characteristics got 'copied' to hers. (This affinity to get other's characteristics into yours is also another characteristic; but then I would be going into too much detail at this stage)
With the knowledge of
1) All current characteristics,
2) The current scenario, and
3) Characteristics and current scenarios of everyone she is in contact with,
We can predict the immediate future;
We can predict which of her characteristics will change;
We can take the new base scenario and new characteristics and predict the next immediate future.
And so, we can predict her life and the life of every human on the planet.
Of course, she can willfully change some of her own characteristics. Like, I can get up one day and decide to walk the stairs everyday from there on. This one change will affect the future. But, with this new data, we can recalculate and predict again.
The future is in our hands and it’s dynamic.
All of this IS complicated. But it is not impossible.
It is a mathematical possibility just as being able to count to a googol is possible. The only thing stopping us is computing power and data collection.
2 Comments:
If you wish to embark on this journey, the first thing you should learn to do is to weed out unrelated data. For eg: in the staircase v/s elevator thingie, how did VB's preference matter? :-P
It is ironical that I just came from here which can be interpreted here to read as follows:
Everything is infinite. Attempts to turn infinite things into finite leads to predictability.
If you are in the future prediction business; I've give you a long standing idea of mine: take a screen canvas of size 800*600px. Assume 8BPP RGB. Generate all possible images (this is surely is finite). You can see every possible event in the universe.
Fasssssscinating post. The implications are truly mind-boggling. Do you know if anyone's trying/tried this? Any research/experiments being done?
I could, however, find two possible obstacles(?) for such an predictive model:
(1) Including the environment as a parameter: I would imagine that a large number of our actions are influenced by the environment - an earthquake would be a dramatic example. Of course, parameterizing the environment would possibly lead to Laplace's dilemma itself.
(2) Human characteristics that we don't completely understand: creativity, passion, ambition, fun, anger, obsession-about-Porsches and a whole host of other seemingly 'non-computable' things.
Hidden Markov Models are somethings that comes to mind when formulating prediction models (usually for phenomena with far fewer characteristics though). I've got a gut feel that your experiment might lead down that path. But of course, HMMs are usually buried in some inane 'will the burglar alarm go off?' situations. Nothing as remarkable as predicting human behaviour or the state of the universe.
Post a Comment
<< Home