Wilhelm Berghammer retweetledi
Wilhelm Berghammer
5 posts

Wilhelm Berghammer retweetledi

A Diffusion Model Framework for Unsupervised Neural Combinatorial Optimization ift.tt/CW0vu4Y
English
Wilhelm Berghammer retweetledi

Autoregressive methods are not only an attractive choice when it comes to language but also for Combinatorial Optimization #NeurIPS2023
Joint work with @SebSanokowski @willberghammer @HochreiterSepp
👇👇👇
Sebastian@SebSanokowski
Our latest paper challenges the status quo in Unsupervised Learning for Combinatorial Optimization (CO) problems.💡 🔄 Unveiling the limitations of popular Mean Field methods, we showcase the power of Autoregressive Approximations of solving CO Problems on graphs. 🧵
English
Wilhelm Berghammer retweetledi
Wilhelm Berghammer retweetledi

You only guess once: We approximate Ising ground states at inference time for Combinatorial Optimization. Drop by at our poster at the #ML4PS2022 work shop!
Time: 4-5 pm
Room: 275-277
ml4physicalsciences.github.io/2022/files/Neu…
#NeurIPS2022
@willberghammer J.Kofler @HochreiterSepp @sebaleh
English