ARC Benchmark Origins
Vložit
- čas přidán 10. 06. 2024
- ARC Prize is a $1,000,000+ public competition to beat and open source a solution to the ARC-AGI benchmark.
Hosted by Mike Knoop (Co-founder, Zapier) and François Chollet (Creator of ARC-AGI, Keras).
--
Website: arcprize.org/
Twitter/X: / arcprize
Newsletter: Signup @ arcprize.org/
Discord: / discord
Try your first ARC-AGI tasks: arcprize.org/play
Yeah I'm excited about this, but isn't it likely that the best solution to ARC will be to treat it like a specialized game like chess? I'm not sure you need intelligence to do well on the problem set. It's very well defined. I can see doing pretty well on the examples by treating each problem category independently. if(moving problem) do moving solution. if(filling out problem) do filling out pattern. If(color change problem) do color change solution. if(subset problem) take the subset. etc.
the problem is that "moving problem" or "filling out problem" are programs/techniques you see in the training data, on the test set there are programs/techniques that haven't been seen before
So your model would need to adapt to those
1N73LL1G3NC3 15 7H3 4B1l17Y 70 4D4P7 70 CH4NG3. -- 573PH3N H4WK1NG leads to ARC-AGI, Th(X/C) guys!
Genius is the ability to create change -- Me if no one else wrote it! (too lazy to l33tcode) hopefully will lead to ARC-AGG, Merci par avance les mecs! :D
Bis because of the quotation! :D