Nabin k Malakar
We are not being able to make objects with artificial intelligence because of the fact that we want to make camera but not associate it with the way our eye works. When we "see" an apple, a lot of logical connections are opened. Camera can take the picture of an apple but it can not associate with its use. When we see a car, we can immediately think of its use. Similarly we can take a picture of an apple. But again, the fact that we have taken picture of an apple will never be reflected in its use by camera. For a robot with camera, apple is an object. Which is not much different than another apple. Just think about an apple, how will you explain it to a robot?
For a record player, one song is no different than another. Yes, you can add the analyzer in terms of currents and voltages and spikes those go up and down and some reference of time in it. But can it sense the melody and harmony of the song? What makes you feel its difference? What makes intelligence to create music?
Similarly, we can have a lot of words in the dictionary. How does dictionary make us understand a word with the help of another word? Without an example of its usage, another word is just as confusing as the first one. During the learning process we start with associating the meanings of words. Association can be with the exact meaning or the nearest neighborhood of the meaning of the subject under construction.
When a robot is given with few objects, it needs to learn their use so that it can learn about the objects. The question here is robot can use the object, but can it learn how to use it? A simple example: Can a robot equipped with first kind of lever go, explore and discover the second or third kind of lever? For this purpose, it needs to associate the functional value of given objects and think of ways to make its usage. The stated example in itself has a lot of complications.
So, I think the next step in Artificial Intelligence is to make connections. Connections that are useful, which can be associated with the functional value, not as objects.
The sustained growth of social network is another example of how connections work. If anything can be identified as having some functional value, its growth is inevitable.
*********
Thinking abilities
Exploration
Discovery
Usage
Discovery
Thinking abilities
cycle et al.
Inspired by the reading of Rota, Feynmann's interview on joys of discovering things and the movie Gods must be crazy and other A.I. related articles.
[Mr. Malakar is a PhD student of Physics. He regularly writes on his own blog Time]
When we identify the objects we assign them with function. That means our degree of intelligence has to do with its usage. The more efficiently we can make the connections, more intelligent we are pronounced. This explains our dominance in the whole animal kingdom. We can make use of things and think of possible ways of connecting the subjects. We can think! The process of thinking leads us to ability to discover. The tool from Thinking to discovery is exploration. The process of exploration is random, however, we would like to have smart ways to explore.
We are not being able to make objects with artificial intelligence because of the fact that we want to make camera but not associate it with the way our eye works. When we "see" an apple, a lot of logical connections are opened. Camera can take the picture of an apple but it can not associate with its use. When we see a car, we can immediately think of its use. Similarly we can take a picture of an apple. But again, the fact that we have taken picture of an apple will never be reflected in its use by camera. For a robot with camera, apple is an object. Which is not much different than another apple. Just think about an apple, how will you explain it to a robot?
For a record player, one song is no different than another. Yes, you can add the analyzer in terms of currents and voltages and spikes those go up and down and some reference of time in it. But can it sense the melody and harmony of the song? What makes you feel its difference? What makes intelligence to create music?
Similarly, we can have a lot of words in the dictionary. How does dictionary make us understand a word with the help of another word? Without an example of its usage, another word is just as confusing as the first one. During the learning process we start with associating the meanings of words. Association can be with the exact meaning or the nearest neighborhood of the meaning of the subject under construction.
When a robot is given with few objects, it needs to learn their use so that it can learn about the objects. The question here is robot can use the object, but can it learn how to use it? A simple example: Can a robot equipped with first kind of lever go, explore and discover the second or third kind of lever? For this purpose, it needs to associate the functional value of given objects and think of ways to make its usage. The stated example in itself has a lot of complications.
So, I think the next step in Artificial Intelligence is to make connections. Connections that are useful, which can be associated with the functional value, not as objects.
The sustained growth of social network is another example of how connections work. If anything can be identified as having some functional value, its growth is inevitable.
*********
Thinking abilities
Exploration
Discovery
Usage
Discovery
Thinking abilities
cycle et al.
Inspired by the reading of Rota, Feynmann's interview on joys of discovering things and the movie Gods must be crazy and other A.I. related articles.
[Mr. Malakar is a PhD student of Physics. He regularly writes on his own blog Time]
