EurIPS is an endorsed presentation venue for NeurIPS papers

Title is slightly misleading. Turns out, sex-disaggregated analysis showed that the supplement group, only in females, gained 0.59 ± 1.61 kg more lean body mass than the controls (p = 0.04)

Direct link to this study: https://www.mdpi.com/2072-6643/17/6/1081

That's an important clarification! So for half the population, creatine works.

So you know, they are misunderstanding the paper.

"There were no group differences in lean body mass growth following resistance training in females (p = 0.10) or males (p = 0.35)"

As a Ukrainian growing up and still living in Denmark and who has spoken to other Ukrainians working at those farms, I have never heard of “slave like conditions”. I’ve heard of hard manual labour and occasional disagreements with the owners of the farms though

Here is example https://fagbladet3f.dk/nu-faar-han-efterbetalt-loen/

The fact that AlphaFold is not able to predict effects of single mutations is well known in protein research community. For SOTA in this space, see Deepsequence (https://doi.org/10.1101/235655) and ESM (https://doi.org/10.1101/622803)

also it is well-known that scripts you downloaded can be run locally

Next step would be to implement autodiff. If only PostgreSQL had CUDA support.

I tried it on a Radeon VII. It is little hard to get running as it does not work with latest kernel, but otherwise it kind of works with some quirks. One of the quirks is the first epoch of your training is very slow to start compared to Nvidia cards. Also, the speed of training is slow.

Slightly off-topic, but I found M1 to be extremely fast at Gaussian Processes (using GPytorch/Botorch). I'm doing a PhD in Machine Learning for protein research, and for the heavy Bayesian Optimisation loop, the MBP M1 is almost twice as fast as my 16 core Ryzen 3950X "deep learning" machine (both use the CPU).

Am I the only one who is not a fan of their design? I get that there might be a slight reference to the legendary Koss SP-3 from 1958, but come on - they don't have to look THAT antique.

https://coolmaterial.com/roundup/history-of-headphones/

I think they look a bit too sterile for something that people "wear". Very corporate, not very soulful. Especially the big blank aluminum surface. The head band also looks rather plasticky. The british stuff, eg from B&W, has more elegance. Combine with nordic minimalism and maybe some natural materials like leather and it would be more appealing. But I'm sure they are technically superior.