Saturday, March 05, 2016

[zjaiakqd] Learning teaching

A beginner chess player plays some games.

Step 1: Use a computer to discover where the player made mistakes.  This is relatively easy.

Step 2: Use a computer to categorize the different mistakes.  Many ways to do this, possibly unsupervised machine learning.

Step 3: Use a computer to devise the most effective plan for improving.  Which category of mistakes should be tackled first?  The most interesting way to do this is to apply machine learning on the progress of many different people, through their games.  Which category of mistakes did they eliminate in which order?

Applicable to more than just chess, though requires a large data set of other people's actions, there mistakes, and their consequent improvement.

No comments :