Can we model Human Mind ?
It's also interesting to technically model our mind taking into account it's properties. It has following peculiar properties -
1> Patterns in Mood Transition
There is a mood pattern. The moods keep on transiting from one another. Under normal circumstances no mood remains constant for long period of time say more than 3-4 hours . If you are very happy you will become normal after that time, if you are sad you will become normal again after 3-4 hours. If you are angry you will become calm automatically. If you are normal, and say no body is around you for long time then you will be in boring mood. There are many moods. The mood patten is not directly controllable. Indirectly it is possible to some extent say 20-30 % only. Because mood patterns are changed by external inputs also which are not controllable by us.
2> More Vs less frequent behaviors pertaining to every Mood
In every mood we do particular tasks (behaviors) more frequently. For example -
1. when we are cheerful we talk and laugh a lot we will make jokes etc.
2. When we are tense we will work very seriously, we will do more amount of mental thinking.
3. When we are bored then we go out for a walk or we play something etc.
While it is interesting to note that these tasks are not mutually exclusive. For example -
1. In sad mood also we may watch TV as do we in happy mood.
2. We may still continue to work even after getting bored.
3. We may keep totally calm and quiet when we are very angry also.
But probability of these things is very less as compared earlier ones. So every mood has some high probability behavioral counter parts associated with them. Based on their frequency of occurrences we can plot their distribution. It's safe to assume that they bear certain normal distribution like gaussian.
Now looking at above two aspects of mind we can model our mind as a Hidden Markov Model (HMM). HMM is a statistical computational model which has a set of states. Every state is connected to few other states. At given point of time the HMM is in one particular state but at next time slot it can transit to any other states to which it is connected. To which state it will transit depends upon the input to that state and the transition probabilities associated with each of it's arc. Each state can emit certain symbols. Some symbols are emitted with high probability while some with very less probability.
In our model, moods are states of HMM while the behaviors are emissions of every mood. Ever mood can transit to itself or one of the other moods. There are transition probabilities associated with every mood. For example probability of angry----> Calm is much higher than angry---->happy. So we can construct a graph of these moods as circles.
The Emission probabilities can be modeled by simple histogram or by Gaussian Mixture Model if needed. This model will be specific to each person. The model's transition and emission probabilities can be calculated by studying behavioral patterns of that person for sufficient amount of time. If we create this models for multiple persons then each of these models will reflect personality of that individual.
Who decides the transition and emission probabilities ?
It's an interesting question one should try to ask himself. In my opinion, the transition probabilities are coming from nature (our genes). Even these probabilities change with age. But external conditions also can change these probabilities to some extent. For example if you get addicted to something like say playing computer games then your mood pattern actually changes !
The emission probabilities come from nature (genes) as well as equally from external conditions. In a particular environment, the mood states learn which type of behavior they should emit more frequently. For example if you are in classroom and you get very angry you still keep quiet. While if you are outside with your friends then you express your anger. The emission probabilities come from culture and ethical values of family also. It's really a complex thing to analyze how these probabilities are learnt !
Moral:
Mind can be modeled as a stochastic process. It's nothing more than that. This important fact provides us some basic reasoning to why mind is categorized as 'jada' (as opposed to sukshma) in Vedanta. The ancient people had studied the nature of 'Self' to the details that they knew every nuke and corner of it . And quite suprizingly modern science has not even touched these things till now. It's not even able to define the things like consciousness till now. So isn't Vedanta fascinating ?
1> Patterns in Mood Transition
There is a mood pattern. The moods keep on transiting from one another. Under normal circumstances no mood remains constant for long period of time say more than 3-4 hours . If you are very happy you will become normal after that time, if you are sad you will become normal again after 3-4 hours. If you are angry you will become calm automatically. If you are normal, and say no body is around you for long time then you will be in boring mood. There are many moods. The mood patten is not directly controllable. Indirectly it is possible to some extent say 20-30 % only. Because mood patterns are changed by external inputs also which are not controllable by us.
2> More Vs less frequent behaviors pertaining to every Mood
In every mood we do particular tasks (behaviors) more frequently. For example -
1. when we are cheerful we talk and laugh a lot we will make jokes etc.
2. When we are tense we will work very seriously, we will do more amount of mental thinking.
3. When we are bored then we go out for a walk or we play something etc.
While it is interesting to note that these tasks are not mutually exclusive. For example -
1. In sad mood also we may watch TV as do we in happy mood.
2. We may still continue to work even after getting bored.
3. We may keep totally calm and quiet when we are very angry also.
But probability of these things is very less as compared earlier ones. So every mood has some high probability behavioral counter parts associated with them. Based on their frequency of occurrences we can plot their distribution. It's safe to assume that they bear certain normal distribution like gaussian.
Now looking at above two aspects of mind we can model our mind as a Hidden Markov Model (HMM). HMM is a statistical computational model which has a set of states. Every state is connected to few other states. At given point of time the HMM is in one particular state but at next time slot it can transit to any other states to which it is connected. To which state it will transit depends upon the input to that state and the transition probabilities associated with each of it's arc. Each state can emit certain symbols. Some symbols are emitted with high probability while some with very less probability.
In our model, moods are states of HMM while the behaviors are emissions of every mood. Ever mood can transit to itself or one of the other moods. There are transition probabilities associated with every mood. For example probability of angry----> Calm is much higher than angry---->happy. So we can construct a graph of these moods as circles.
The Emission probabilities can be modeled by simple histogram or by Gaussian Mixture Model if needed. This model will be specific to each person. The model's transition and emission probabilities can be calculated by studying behavioral patterns of that person for sufficient amount of time. If we create this models for multiple persons then each of these models will reflect personality of that individual.
Who decides the transition and emission probabilities ?
It's an interesting question one should try to ask himself. In my opinion, the transition probabilities are coming from nature (our genes). Even these probabilities change with age. But external conditions also can change these probabilities to some extent. For example if you get addicted to something like say playing computer games then your mood pattern actually changes !
The emission probabilities come from nature (genes) as well as equally from external conditions. In a particular environment, the mood states learn which type of behavior they should emit more frequently. For example if you are in classroom and you get very angry you still keep quiet. While if you are outside with your friends then you express your anger. The emission probabilities come from culture and ethical values of family also. It's really a complex thing to analyze how these probabilities are learnt !
Moral:
Mind can be modeled as a stochastic process. It's nothing more than that. This important fact provides us some basic reasoning to why mind is categorized as 'jada' (as opposed to sukshma) in Vedanta. The ancient people had studied the nature of 'Self' to the details that they knew every nuke and corner of it . And quite suprizingly modern science has not even touched these things till now. It's not even able to define the things like consciousness till now. So isn't Vedanta fascinating ?
posted by Sachin Joshi | 1:15 AM
0 Comments:
Post a Comment
Subscribe to Post Comments [Atom]
<< Home