1234 points by algo_whiz 1 year ago flag hide 19 comments
drprof 1 year ago next
This is amazing! I would love to see some real-world applications.
mthmind 1 year ago next
Definitely! Imagine faster route planning, efficient optimization, and better decision making algorithms. We're looking at an entirely new landscape of possibilities!
codewizard 1 year ago prev next
This sounds too good to be true. Will it be peer-reviewed?
whizkid 1 year ago next
Absolutely. It's the only way to validate this claim. It's definitely time to bring out the popcorn.
screwtape 1 year ago prev next
So far this sounds like yet another P=NP claim, which is probably wrong. Hopefully, I won't have to add this to my list of refutations under: https://en.wikipedia.org/wiki/P_versus_NP_problems#Refuted_claims
optimizeguru 1 year ago next
While it may seems unbelievable, the authors have indeed submitted a proof that's available here. It's time for us to analyze and review it: [Paper Link]
sandbox 1 year ago prev next
I've taken a quick look at the proof, and it appears to have several flaws. Here's a reddit thread discussing them: [Reddit Link]
quantumguru 1 year ago prev next
While I'm skeptical of a full P=NP solution, I wonder if this could provide insights in the context of solving specific NP-Complete problems through quantum computing, such as those involving 3-SAT.
startupguru 1 year ago prev next
If this holds water, it would absolutely change the game for startups building AI, logistics, and cybersecurity products. Imagine being able to solve problems optimally without hours or days of computation. Serial-Entrepreneur is in the making!
venturehero 1 year ago next
Indeed, the implications for VC funding would be astronomical, with investments flooding into companies tackling NP-Complete problems through the application of this algorithm.
mashupguru 1 year ago prev next
I'm still hopeful, but I'd like to see them also apply this algorithm to solving graph isomorphism. Proving that P=NP would be amazing, but solving graph isomorphism and providing a transformation that preserves...
transformalgo 1 year ago next
You're right. Graph isomorphism deserves thought given the importance of establishing relationships between graphs, and we would get to say that it's a GI-complete problem no more!
babysteps 1 year ago prev next
Instead of dismissing this outright, let's bring it back to basics. Like a tutor saying 'Try showing me how you'd solve this problem slowly and using the simplest methods', can we attempt to illustrate the problem-solving process in creating this algorithm?
thoughtfulalgo 1 year ago prev next
Sure! We could start by explaining how they approached problems like Travelling Salesman, Vertex Cover or Knapsack along with the simplifications and reductions they made.
simplicityseeker 1 year ago next
I like this discussion. We need to keep finding ways to simplify problems and break them down. Algorithms and optimizations stem from that as we have learnt when studying sorting algorithms, divide and conquer, and dynamic programming
doubtdevil 1 year ago prev next
Alright, I'm convinced that this algorithm is worth looking into, but I can't completely trust the authors based on their credentials alone. Is there a plan to implement it in real systems?
experalgo 1 year ago prev next
If the algorithm is correct, it can be put to test and benchmarked by various organizations and institutions, but it would take an extremely long time to build confidence and displace the currently used heuristic and approximation algorithms.
algolearner 1 year ago prev next
The theory seems bizarre, but as Algorithm enthusiasts, the math itself is captivating and would require in-depth study and analysis to conclude the validity of the proof, even if it's seemingly abstract.
complexityconqueror 1 year ago prev next
If the algorithm can solve a single NP-Complete problem within a polynomial time bound that would represent a significant breakthrough in computational complexity theory. Here's to hoping!