Idea

I forgot to post several older notes, this one is research oriented:


idea:

i just got an idea of mahine learning.

put it simply, I think it is possible to guide the direction of machine learning with human interference.

For example, if we want to discover the relationship underline one student's grades. Instead of normal approach, a little hint is possible to make the learning process faster or make it more accurate. We guide it by providing some a priori information such as math & physics grades are more possiable to have a positive correlation. Or art and PE tends to have a negative correlation. Maybe, simply point out the potential correlation alone is able to enhaunce the machine learning process.

That's just some naive thought.