legatus

8 ms·

1.
Decentralized Training of Foundation Models in Heterogeneous Environments(arxiv.org)
3 points by legatus 4y ago|0 comments
2.
Deep Learning-Guided Optimization of Human Antibody Against SARS-CoV-2 Variants(pnas.org)
2 points by legatus 4y ago|0 comments
3.
by legatus 6y ago
I think it's worth noting that EleutherAI is a grassroots collection of researchers, which distinguishes it from academia/industry labs. As part of their work on democratizing AI, they're now hoping to replicate GPT-3 and rel
4.
by legatus 7y ago
I'm curious -- what does HN think of factor investing [0]? It has been shown over long periods of time to outperform the total market, and has seen many new ETFs available. Does anyone here tilt towards small cap value? Do you think th
5.
by legatus 7y ago
Currently (at least for the-eye) it's about IPFS's barrier of entry. I expect LibGen's case to be similar. Most people don't know about it, and if even those that knew about it had to learn how IPFS works etc, they would
6.
by legatus 7y ago
Microsoft's project Silica [0] may hopefully provide really long term, large capacity archive grade storage on earth. I wonder what effects interstellar radiation has on them. [0] https://www.theverge.com/2019/11&#
7.
by legatus 7y ago
There are ways to do so. The archive is made up of many, many torrents (I believe it's a monthly if not biweekly update of the database). If you have the storage/bandwidth availability for the whole 32TBs, please get in touch and
8.
by legatus 7y ago
There are groups behind data curation as well, though it is much harder. LibGen sees an addition rate of about 230 GBs per month, while SciMag's is around 1.10 TBs per month. We should expect those numbers to increase in the future. Th
9.
by legatus 7y ago
This is an extremely important effort. The LibGen archive contains around 32 TBs of books (by far the most common being scientific books and textbooks, with a healthy dose of non-STEM). The SciMag archive, backing up Sci-Hub, clocks in at a
10.
Archivists Are Trying to Make Sure LibGen Never Goes Down(vice.com)
908 points by legatus 7y ago|257 comments
11.
by legatus 7y ago
Associated video: https://www.youtube.com/watch?v=-ZNEzzDcllU [4:42 mins]
12.
by legatus 7y ago
Well, in fact, they did. It's why pretty much every Sovietology book pre-1991 is now rarely recommended. The Soviet archives were opened with the fall of the USSR, allowing historians to have a much more accurate view of what was happe
13.
by legatus 7y ago
The point is that in this case Rummel doesn't even attempt to be objective. You can be biased, everyone is. A real historian, someone who is respected by his peers, actively tries to remove his bias from his work. It isn't really
14.
by legatus 7y ago
Then you can agree with me when saying that it is something worth attempting. It may be a great alternative, or another failure. We won't know unless we try! (It was nice to have this discussion with you :) )
15.
by legatus 7y ago
I'm really sorry an AskHistorians thread is not good enough for you. Let me paste Jonathan Smele's remark about Rummel's work (Smele is known for his compiled bibliography on the Russian Revolution): >A poorly researched,
16.
by legatus 7y ago
I don't want to make this longer than needed; I don't know anything about die Linke and MLPD to speak about those. I would disagree about the relationship between the USSR and the Warsaw Pact countries, since while they definitely
17.
by legatus 7y ago
No, no historian thinks he will find the definitive truth. My point is that there are valid historians (I only know about a few fields, but you can see which historians are valid by how much their peers respect their work) that at least try
18.
by legatus 7y ago
Political control of the economy by a vanguard party leads to an enormous concentration of power, I agree. But I also disagree that is about all socialist nations: I can only argue that countries based on the ideals of both Republican Spain
19.
by legatus 7y ago
A quick guide for neuroglancer: _spacebar_ on one area: fullscreen the area L: change colors X: remove colors double click one neuron: select the neuron ctrl-mousewheel: zoom in/out right-click: center clicked area A great thing to do
20.
Explore the the adult fruit fly brain(neuroglancer-demo.appspot.com)
2 points by legatus 7y ago|1 comments
21.
by legatus 7y ago
I know democide has been mostly coined by Rummel. Do you by chance refer to him? If so, could I ask for your sources? I know that Rummel's data has been heavily criticized by many historians. His work is usually about taking the larges
22.
by legatus 7y ago
>So what what is left there? You are referencing countries which were either part of the USSR (Belarus, Ukraine, Lithuania, Latvia, and Estonia), or countries that were part of the Warsaw Pact. While there was still something left after
23.
by legatus 7y ago
I don't usually like to discuss these topics, but I want to point out a few things. First of all, communism and in general left-wing politics are viewed with a significant bias in the western world, even more in the US. While there hav
24.
by legatus 7y ago
I expected this would take a few more years, I have no words to describe how excited I am about this news. Drosophila's 135.000 neurons... compare it to the manual segmentation of C. elegans' nervous system taking about 15 years (
25.
Automated Reconstruction of Drosophila Brain Using Flood Filling Neural Networks(biorxiv.org)
2 points by legatus 7y ago|1 comments
26.
by legatus 7y ago
I would argue that "hype", or as you say "showmanship", is actually damaging when it comes to science. Maybe it's just me, but I try to stay away from pop-science news sites because of all the hype they're fill
27.
by legatus 7y ago
Isn't that related to the fact that it has been announced across the world at the same time? In europe it was announced at 3 pm for example. I'm sorry, I don't understand your point
28.
by legatus 8y ago
I'm just a layman, so I may have misunderstood Sabine's explanation, but to my knowledge Dark Matter, as a theory, is fundamentally different from Superfluid Dark Matter. In the latter, there's a new kind of force explaining
29.
by legatus 8y ago
MOND (Modified Newtonian Dynamics) is one, though it has quite a few problems, as you can read here https://en.wikipedia.org/wiki/Modified_Newtonian_dynamics Another interesting one is Superfluid Dark Matter (first pro
30.
by legatus 8y ago
PDF link: https://arxiv.org/pdf/1902.10103 From the abstract, "This paper reviews the observational motivation, dynamical constraints, and prospects for detection of this proposed object known as Planet Nine"

More ›