Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

This is interesting. My intuition says NNs will be terrible at combinatorial optimization types of problems. Then again Go is a combinatorial type of game and AlphaGo beat everyone.

Maybe an NN for Sudoku should use AlphaGo type of architecture.



Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: