Very interesting, but still in the line of [1.human solves the puzzle] >> [2. implement mannualy]. The big challenge is here hidden in 1 and 2: build a system that deduces/builds/selects the approach, and then write the specific set of instructions to satisfy the task.
Very cool and detailed notebook about using cellular automata as a DSL for generating ARC solution programs.https://www.kaggle.com/arsenynerinovsky/cellular-automata-as-a-language-for-reasoning …
-
-
-
From the biggest univerality (basic priors or meta priors) to the smallest specificity (THIS task).
- Show replies
New conversation -
-
-
Great notebook indeed! Still busy getting a better idea of how to generate cellular automata iteratively. Added some ideas: https://www.kaggle.com/c/abstraction-and-reasoning-challenge/discussion/131195 … CA seem like a capable representation, but I'm a bit afraid of the "weak causality" regarding minor changes to their rule sets.
-
This is killing me :-) going to bed at 2, 3 am every night, I need to work!pic.twitter.com/D8swqcAzLB
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.