Anyone studying Computational Intelligence (AI) ?
Posted by 33957800
| Announcements | Last Post | |
|---|---|---|
 |
SoC Curricula | 09/30/2017 01:08PM |
|
Demarcation or scoping of examinations and assessment | 02/13/2017 07:59AM |
|
School of Computing Short Learning Programmes | 11/24/2014 08:37AM |
|
Unisa contact information | 07/28/2011 01:28PM |
|
Re: Anyone studying Computational Intelligence (AI) ? February 27, 2006 09:38AM |
My Background is actually in philosophy.
My goal is to pursue research into the foundations of cognitive science, which in a narrow sense is heavily reliant on the "computational account" of mind. My problem was that I was coming up against a lot of really heavy theory (at least what I consider heavy) regarding the limits of computation and the completeness of physics. It's always the maths that kills me
With your background though, I doubt you'll find much of it difficult (the cognitive science/philosophy) Even though some of the postmodern critiques of mind get a little tricky.
Your research sounds fascinating. Do you have anything online or published that I could take a look at?
I think you will enjoy 3rd year logic as well, there is some useful stuff in that module regarding defeasible reasoning (heuristics) etc. That you would probably find super interesting.
Applied logic is where I hope to eventually end up doing work, although there doesn't seem to be a lot of support for it in SA (academically speaking)
My goal is to pursue research into the foundations of cognitive science, which in a narrow sense is heavily reliant on the "computational account" of mind. My problem was that I was coming up against a lot of really heavy theory (at least what I consider heavy) regarding the limits of computation and the completeness of physics. It's always the maths that kills me
With your background though, I doubt you'll find much of it difficult (the cognitive science/philosophy)
Even though some of the postmodern critiques of mind get a little tricky.
Your research sounds fascinating. Do you have anything online or published that I could take a look at?
I think you will enjoy 3rd year logic as well, there is some useful stuff in that module regarding defeasible reasoning (heuristics) etc. That you would probably find super interesting.
Applied logic is where I hope to eventually end up doing work, although there doesn't seem to be a lot of support for it in SA (academically speaking)
|
Re: Anyone studying Computational Intelligence (AI) ? February 27, 2006 02:36PM |
Nothing online yet, I have two small black notebooks scribbled full of it and have to put it into a formatted publication first as well as continously review it for soundness (as I gain more background into formal logic and cognitive processes) - ie critically look at my statements/reasonings to correct the falacies and inductive reasonings mistaken for truths - its difficult to avoid affirming a notion as oppose to objectively deducing a true conclusion
However, so far I can say that once the problem has been solved of finding a sound algorithm to implement the contextualization of concepts the result will be something that needs to be 'brought up' like a new born baby for at least 4 years and then duplicated and exposed to the clones who will then grow each others reasoning capabilities by interaction between themselves and multiple humans from different cultures and backgrounds (to avoid the same problem I mentioned above wrt decerning informal logic conclusions from formal logic conclusions). To avoid a threshold (stale condition) where they all know and understand the same and therefore will be unable to continue 'evolving' their capacity for sound reasoning, there has to be a measure of randomness in each. Pseudo randomness will not be sufficient and therefore a very sound means of producing randomness will have to be invented.
The randomness might proof to be more difficult than the contextualization problem - effectively randomness means one 'seed' must be as non-linear as the definition of non-linearity is (each element has exactly the same probability as the next of occuring) and this is very difficult to achieve in a linear discreet system such as software (which essentially is only a configuration sequence of a configurable state machine which we call a computer - another example of a state machine - not configurable - would be a traffic light, it has at least 7 states: green - wait - amber - wait - red - wait and if undefined state is reached then flash red light).
It sounds interesting, the angle from which you come, I would look at the possibility of furthering academic work abroad if such a field is not in demand in South Africa. For example I doubt there is a market for AI programmers in SA (game development doesnt really need high quality AI, just a good approximation to mimick real response and meaningful unpredictable behaviour and Denell would probably not be at the point of introducing AI into weapons guidance and threat avoidance systems). The US would seem a more likely country for this and your kind of work as they invest more into such concepts that might proof valuable in the future. Besides there you have government, civil and academic organizations that see the value of such work. (I dont know of european or eastern countries, havent had much information about them - however I think Japan for example would rather value home grown intellectual property than introducing it from abroad).
However, so far I can say that once the problem has been solved of finding a sound algorithm to implement the contextualization of concepts the result will be something that needs to be 'brought up' like a new born baby for at least 4 years and then duplicated and exposed to the clones who will then grow each others reasoning capabilities by interaction between themselves and multiple humans from different cultures and backgrounds (to avoid the same problem I mentioned above wrt decerning informal logic conclusions from formal logic conclusions). To avoid a threshold (stale condition) where they all know and understand the same and therefore will be unable to continue 'evolving' their capacity for sound reasoning, there has to be a measure of randomness in each. Pseudo randomness will not be sufficient and therefore a very sound means of producing randomness will have to be invented.
The randomness might proof to be more difficult than the contextualization problem - effectively randomness means one 'seed' must be as non-linear as the definition of non-linearity is (each element has exactly the same probability as the next of occuring) and this is very difficult to achieve in a linear discreet system such as software (which essentially is only a configuration sequence of a configurable state machine which we call a computer - another example of a state machine - not configurable - would be a traffic light, it has at least 7 states: green - wait - amber - wait - red - wait and if undefined state is reached then flash red light).
It sounds interesting, the angle from which you come, I would look at the possibility of furthering academic work abroad if such a field is not in demand in South Africa. For example I doubt there is a market for AI programmers in SA (game development doesnt really need high quality AI, just a good approximation to mimick real response and meaningful unpredictable behaviour and Denell would probably not be at the point of introducing AI into weapons guidance and threat avoidance systems). The US would seem a more likely country for this and your kind of work as they invest more into such concepts that might proof valuable in the future. Besides there you have government, civil and academic organizations that see the value of such work. (I dont know of european or eastern countries, havent had much information about them - however I think Japan for example would rather value home grown intellectual property than introducing it from abroad).
Sorry, only registered users may post in this forum.