Showing posts with label Concept. Show all posts
Showing posts with label Concept. Show all posts
One of the key reasons why my model inspires me is its predictive potential ! Unlike other psychological theories which are based on statistical data, we use a model which is built on a set of verifiable postulates from which the rest of the predictions are derived. It is here that we have an edge over conventional psychologists who think human behaviour is too complex to be modelled by numbers and functions.
Reminds me of the situation in biology before Darwin, when all that was possible was to observe and record data about the innumerable kinds of plants and animals our planet hosts. But once we had a theory built on verifiable postulates - the theory of natural selection, we were able to predict what kind of organisms survive an evolutionary phase in a particular environment and hence predict what kind of beings might inhabit an ecosystem.
Also reminds me of the situation in economics, national planning and warfare before game theory. When all you could do was to send out your best orders and hope that the opponent does something that will benefit you. But once we had the simplest of framework in hand, we were able to predict how a market would behave, how nations make decisions and so on.
And so, we make predictions, verifiable statements about learning, altruism, decision making time, dream frequency, laziness and a hundred other things that form a part of our everyday lives ! Most of them are already known observations and hence the so called predictions turn out to be explanations. And the remaining few, we hope will be found true. But what strikes me is that all of these statistically/inductively known facts are now reasoned out by using a minimal set of postulates on which we have built out model. That, I see as the biggest boost to my confidence in the model; like we say in physics - "The theory was so beautiful that it had to be true."
Reminds me of the situation in biology before Darwin, when all that was possible was to observe and record data about the innumerable kinds of plants and animals our planet hosts. But once we had a theory built on verifiable postulates - the theory of natural selection, we were able to predict what kind of organisms survive an evolutionary phase in a particular environment and hence predict what kind of beings might inhabit an ecosystem.
Also reminds me of the situation in economics, national planning and warfare before game theory. When all you could do was to send out your best orders and hope that the opponent does something that will benefit you. But once we had the simplest of framework in hand, we were able to predict how a market would behave, how nations make decisions and so on.
And so, we make predictions, verifiable statements about learning, altruism, decision making time, dream frequency, laziness and a hundred other things that form a part of our everyday lives ! Most of them are already known observations and hence the so called predictions turn out to be explanations. And the remaining few, we hope will be found true. But what strikes me is that all of these statistically/inductively known facts are now reasoned out by using a minimal set of postulates on which we have built out model. That, I see as the biggest boost to my confidence in the model; like we say in physics - "The theory was so beautiful that it had to be true."
I remember saying that it all started with a crazy idea... and 3 months later, this is where the idea stood -
We were working on developing a software to create a virtual personality of the user so that the same can be loaded to a different environment to predict the user's reaction to the situations. This would have a lot of applications in various fields which require HR assessment and development... (Yea, I'm being hopelessly modest - it's just one small target customer base).
The problem however, was to work out a model which would allow us to arrive at such an accurate model of the user's personality since most of predictive psychological theories are accurate only on large population groups.
We were basing our model more on evolutionary biology & game theory, with psychology to estimate certain aspects of the model rather than using it as a starting point. We were working with help from a few scientists in evolutionary biology & psychology from IISc, NIMHANS, NBRC & Oxford regarding the same.
We had completed an initial model for an individual's personality and were looking forward to getting out a design document for the software so that we can atleast pitch the idea to a professional AI programmer. But the big question that we asked ourselves a zillion times was - "How can we get the best team for the project?!"
We were working on developing a software to create a virtual personality of the user so that the same can be loaded to a different environment to predict the user's reaction to the situations. This would have a lot of applications in various fields which require HR assessment and development... (Yea, I'm being hopelessly modest - it's just one small target customer base).
The problem however, was to work out a model which would allow us to arrive at such an accurate model of the user's personality since most of predictive psychological theories are accurate only on large population groups.
We were basing our model more on evolutionary biology & game theory, with psychology to estimate certain aspects of the model rather than using it as a starting point. We were working with help from a few scientists in evolutionary biology & psychology from IISc, NIMHANS, NBRC & Oxford regarding the same.
We had completed an initial model for an individual's personality and were looking forward to getting out a design document for the software so that we can atleast pitch the idea to a professional AI programmer. But the big question that we asked ourselves a zillion times was - "How can we get the best team for the project?!"