Wednesday, July 12, 2006

Machine Learning Continued! - 2

Hey Guys,

I am back. Well, its after a long gap. But I hope now I am more consistent.

Thanks to Mohit for his insightful comment. He mentioned some very cool points ->
  • We choose to learn or not from experience.
  • We learn from our and others' mistakes.
I believe a more intelligent human being will choose to learn from experience. Though I don't have a formal proof. So, taking this assumption, lets go ahead with building a machine which learns from experience. In absence of this way of learning, I don't know how can a machine become intelligent. Thoughts are invited in this line of thinking.

The second point is pretty cool. But I guess this also comes under experience only. This is more of a reward/punishment procedure. So, father says to the child -> if you sit quietly I give you a chocolate.

If chocolate is defined as a reward, then sitting quietly might be a fruitful activity in the future, in case the machine is looking for a reward. Keeping in mind that there might be situations in which we are looking for punishment. :~)

So, guys I think we can march ahead with this exiciting discussion.

Please pour down your ideas.

Have a great time...

No comments: