Lately I have been consumed in the learning of the concepts
and algorithms for Machine Learning – a branch of Artificial Intelligence. While
dueling the ideas and algorithms and the magical mathematical notations that
allow us to understand the existing patterns and enabling the predictions, a
very interesting thought came to observation.
Most of the prediction algorithms required us to understand
the test scenarios and then create clusters based on measurable parameters.
This information and the clustering algorithm trains the system or in other
words develops a stereotype in the minds of the observer. Based on the
stereotypes, the observers start observing and then deciphering every
observation henceforth.
While the strategy works amazingly well when it comes to
innate objects, I see a big trouble in extending the same to the thoughts
getting generated in an individual. No thoughts in an individual are discreet
and are neither mutually exclusive so that we can apply them to a clustering
algorithm. The thought process in an individual behaves both as a continuous as
well as discontinuous function based on the variety of factors including the
time and the thought itself. It is so very difficult to map it on a graph and
mimic a human being.
Extrapolating the individual brain to a group or a whole
region, we are looking at a much larger diverse dataset. Applying a clustering
algorithm which is the most optimum fitting for the dataset would rather be
either overfitting for some and under fitting for others. There shall never be
a fully fitting algorithm that works all the time.
Think of the interpersonal interactions we have on a day
basis. Every time, we meet a new person, we try to judge them – classify them
based on the stereotypes we have developed in our mind. Based on the
classification model that our brain reports, our interaction is molded
accordingly. Based on this classification, we place the individual in a
referential frame with ourselves and then we interpret every conversation and
respond accordingly. When talking to a group of people, the same strategy
applies.
When dealing with a community, region or a large amount of
people a well-oiled administration tries to divide the groups into smaller groups
so that it can find commonalities and leave out the rest. What happens to those
who would fall partly in multiple groups?
Now the question arises again from an individual’s
perspective – what happens when you feel that you are not being heard or you
are being misinterpreted? The may even believe that their identity is being
destroyed since they are misinterpreted. The notion of being misinterpreted
causes emotional imbalance in an individual. The individual (which could be me
or even you) starts feeling that they are being shunned. By denying someone
their thoughts or ideas or even identity, this is a form of violence –
emotional violence. When this individual is again extrapolated to a community
or group, the violence feeling is also expanded manifolds. Imagine, when the
emotional violence reaches a threshold and starts erupting? It may take forms
of physical violence and many other ugly forms?
Administration or grouping or clustering is thus the cause
of violence in us. When we can understand this concept, and blame it all on the
administration techniques employed by the politicians, leaders and governments,
why shouldn’t we question ourselves? We must first demolish the habit of
building up stereotypes in our head and be open in our communication.
It is easy to handle the individual’s sense of violence though
empathy. However, if it grows as a general thinking, it may be an unstoppable
phenomenon.
Artificial Intelligence is great – but it is artificial – no
match to your own intelligence.
Comments
Post a Comment