Learning From Animations

While playing with Many Faces of Go, I have activated the option to view the moves considered by the program.

The result is a fast animation of possible local developments.

While it's to fast to be meaningfull for a human, one can still recognize common shapes.

I wonder if a specially crafted animation could be used to inprint in to the visual memory of the subject possible developements of common shapes, josekis and tesuji.

The teaching process could be enhanced further with special effects like

• spectacular death of captured groups
• special highlight of groups that become unconditionally alive
• highlight of terrtory under influence

-- HET

Karl Knechtel: A fascinating idea. I'm not sure it would work, though. I imagine that more time is spent considering bad moves (because the vast majority of moves are bad, and the computer can't mark a move as bad until it finds a good refutation, where it will try bad ones first... so on recursively to the limits of its reading depth). Although it would be interesting to see ladders read out, for example ("oh, it *does" work; that's why the program chickened out...").

I've started a topic about this on the discussion page

And another about choosing a name

--HET

Links

This article seems interesting in our perspective:
Learning to discriminate complex movements: Biological versus artificial trajectories
BTW The whole site deserves digging.