Mateus Mota (@mateuspontesm) header
Mateus Mota (@mateuspontesm) profile picture
Joined June 2017
Photo by Mateus Mota (@mateuspontesm)Photo by Mateus Mota (@mateuspontesm)Photo by Mateus Mota (@mateuspontesm)Photo by Mateus Mota (@mateuspontesm)Photo by Mateus Mota (@mateuspontesm)Photo by Mateus Mota (@mateuspontesm)
The NeurIPS 2023 Neural MMO Challenge is now live on AICrowd! Train agents that generalize to new tasks, maps, and opponents. Academic tracks are limited to 8 GPU hours to prevent resources from dominating.
5
28
2
143
31,433
I packed-up a full-text paper scraper, vector database, and LLM into a CLI to answer questions from only highly-cited peer-reviewed papers. Feels unreal to be able instantly get answers by an LLM "reading" dozens of papers. 1/2
40
309
44
2,312
525,624
Do you have any idea how much time something like this would have saved me during my PhD... This add-on displays tabs as a hierarchical *tree* rather than a flat list. Extremely useful for research! github.com/mbnuqw/sidebery
5
16
1
126
TRÊS MITOS SOBRE A ECONOMIA BRASILEIRA: A QUEM SERVEM? (vou dividir em 4 partes, um fio por dia) Parte 1 1. Desde 2015, a sociedade vem sendo submetida à narrativa de que o dinheiro acabou e a austeridade fiscal é a única saída para a retomada econômica e o controle de preços
74
1,078
68
6,097
Mateus Mota retweeted
For AI/ML people curious about Mastodon — check out sigmoid.social! It’s fast-growing, focuses on AI, and has a good harassment policy. It’s run by @gradientpub, so I have hopes it will be managed well long-term. You can find me there too: sigmoid.social/@lowd
5
10
2
81
Mateus Mota retweeted
A milestone result in AI for science & for biology. The first large scale map of metagenomic protein structures. >600 million proteins, far larger than anything before. This will have far reaching implications for our understanding of biology and the building blocks of life.
Announcing the ESM Metagenomic Atlas — the first comprehensive view of the ‘dark matter’ of the protein universe. Made possible by ESMFold, a new breakthrough model for protein folding from Meta AI. More in our new blog ➡️ https://teapawt.net/t.co/LsUhSjzqCf 1/3
12
96
6
537
"What Makes Convolutional Models Great on Long Sequence Modeling?" CNNs—not transformers—now dominate the hardest sequence modeling benchmark. Here's how this happened: [1/14]
31
399
35
2,005
Amazing. VSCode LaTeX Workshop has dark mode for *PDFs*! It even inverts the figures! github.com/James-Yu/LaTeX-Wo…
9
125
7
1,079
Mateus Mota retweeted
Do you want to understand how Diffusion Models work? Stop searching on Google. Here you have one video, one article, and one paper that will tell you everything you need to know: 1 of 5
65
574
11
3,446
Name a more iconic duo than paper deadlines and \vspace{-1em}
6
4
96
Mateus Mota retweeted
10 free websites so useful they should come pre-bookmarked on every browser:
1,132
15,579
664
73,672
“A still of Kermit The Frog in Blade Runner 2049 (2017)” #dalle
107
1,723
1,226
8,168
Mateus Mota retweeted
Last year PhD student applying for jobs
Nature editor and his assistant desk rejecting papers
1
2
58
Second order effects of the rise of large language models:
70
635
195
2,940
Mateus Mota retweeted
Dear 2022, Thank you for submitting your manuscript. I regret to advise that I did not feel as enthusiastic about it as I wanted to. First of all, there are simply too many plot lines for the average reader to follow.
83
1,689
306
8,322
Pandas Tutor lets you write Python pandas code in your browser and see how it transforms your data step-by-step. pandastutor.com
4
137
5
737
I am blown away by mathcha.io. Such a useful tool for research. This in-browser graphical LaTeX tool gives you free-form drawing (tikz export), WYSIWYG rendering, symbol shortcuts, and even picture-based symbol search. I might even write full papers in this...
7
124
6
833
Mateus Mota retweeted
Implicit Q-learning, or "I can't believe it's not SARSA": state-of-the-art offline RL results, fast and easy to implement; almost SARSA, but with a different loss to provide "implicit policy improvement": arxiv.org/abs/2110.06169 w/ @ikostrikov, @ashvinair 🧵->
2
23
2
118