Machine learning is undoubtedly a great field. But I cannot stop quoting it as an absolute GENERALIZATION procedure where we try to extrapolate based on the past. It works miracles because, people predominantly do things in a stereotypical manner which allows machines to predict our next moves.
People mostly fall under this category. But, what if we – the humans start out-of-the-box thinking? If everybody tries deceptive thought processes, then machine learning applied to those aspects of life, where it’s not bound by the laws of nature?! What would it be like.!?!? Haha.. I feel you. I could just imagine, what I tried making you imagine and traverse your state of mind through. Even now, if I have succeeded making your thoughts traverse through my paths, it is because of the foreseeable nature of our thoughts which are based on the common belief systems around what’s right/wrong, good/bad generally and acceptably. For instance, banging ones head against the wall is painful and we all agree on that. However, if an unsupervised model is launched to learn its path without obstacles and is used to help people with disabilities around making cognitive judgements while they move, it would be an almost 100% success with an exception being for those individuals that do not think banging their head to a wall is painful haha. And, these are the anomalies (according to the model but are normal according to the practitioners of this belief) that can deceive the model and get to stay on the other side of the wall!! The OUT-THINKERS. Machine learning is got to be wary and worried about these data points!!
Machine learning is guaranteed to work on those aspects impeccably – wherein we make our machines predict based on the physical laws (laws of motion etc.). For instance where we implement AI ( Artificial Intelligence) for cars to park themselves based on the available space around, auto pilot of vehicles, driver-less cars etc. These acts include physical, non-living or immobile entities which are not autonomous except during extremities such as heavy wind which could actuate them with some kinetic or potential energy. Otherwise remain predictable.
But, what we’re trying to use machine learning or AI of late, is to predict the human thought process or homo sapiens’ instincts or reactions to a certain action based on their previous reactions to similar actions. However, this works only as long the situation makes us feel the same at both the points in time.
1. When I first faced this situation
2. Second or the consecutive iterations, when I encountered a similar situation.
For instance, Let’s take my situation.
I don’t like curd rice. So first time I would deny to have or even puke if forced to have.
But the 3rd or 4th or even the 2nd time when I had, irrespective of the events between the two points in time and the length of the interval between the two attempts under consideration, I might start liking it for no reason whatsoever. Even possibly start having it ravenously. So if we try substantiating by saying that the machine learning model would learn from this act and turn its end class variable positive wrt my reaction to curd rice, then how sure am I to not start hating curd rice again. LOL
This is a very simple but a subtle example to just show that, a similar or even the same action might not stimulate a human being to give out the same old reaction as is previously.
For instance if the Indians who are well known for their savings, stop saving and start spending due to the betterment of the government services to people, then the entire banking sector of the world’s so called largest democracy will be stumped down in no time.
After all, Machine learning is to give a machine its autonomy to just try and emulate its creator’s species. LOL
We shouldn’t be surprised if it fails.
Are we running short of people that we’re running behind machines to emulate humans.? Why are we setting a timer for our self-destruction. Of course, machine learning should be used. Where? There, where human intervention is harmful. In the space, fire, under water, in extremities and not in our living rooms and kitchens and bathrooms, where we at some point in time will evolve (in fact devolve) to delegate our relationships to them to be handled with care ROFL.
Rather than training the machines, why not spend a part of those resources on the training of human beings in providing them the basic education. No matter how far we’d go in this spree of innovation, but would we even be able to reach the half mark of god’s creation? Which is our self.!! In fact Billions. Sorry. Zillions of dollars are being dumped in with the hope of creating the copy of the 1500 grams of grey matter or even something that works 10 to 20 % similar to that when we already have billions of such grey matters untrained (rather say unused). When god has already given us the power of giving birth to new offspring who can be tamed to excellence, but are being left out thinking machines are far better options relatively; whom are we fooling. Not OURSELVES?
P.S. I have enrolled for a machine learning course. Hahaha.
Please excuse me if you happened to find any grammatical errors. Cuz I have not got this article proof read. I have just written down my flow of thoughts as is and free. Any kinda comments showered are accepted and are appreciated.
Absolutely no offense to machine learning on the whole, but to those aspects which we are trying to over do.
Appreciate your time 🙂
Thoughts Flow 