PyGeneses is a Deep Reinforcement Learning framework that attempts to simulate artificial agents in bio-inspired environments. One of the use cases features researching various possible behavior trends and drawing parallels with the real world. https://t.co/sfNC5WJsy8
— PyTorch (@PyTorch) December 4, 2020
My unsolicited advice to academic job seekers: if there is a department that keeps losing faculty, don't believe that the departures are random. The place is a shit show; run like hell.
— KLDivergence (@KLdivergence) December 4, 2020
One Learning to RL them all:
— Yann LeCun (@ylecun) December 4, 2020
ReBeL (Recursive Belief-based Learning) is a general RL+Search method that works for all two-player zero-sum games, including imperfect-information games (poker, liar's dice,...) and perfect-information games (chess, go....). https://t.co/2sw8Zbe8rg
Native. Pipe. In. Base. R.
— Colin Fay 🤘 (@_ColinFay) December 4, 2020
🎉 🎉 🎉 #RStats https://t.co/0hoXlixtdg
Materials for my class "beyond gradient descent". After carefully explaining the different types of convergence rates (rarely explained), I cover coordinate descent, Newton's method, Frank-Wolfe and mirror descent.
— Mathieu Blondel (@mblondel_ml) December 3, 2020
Slides: https://t.co/4Pk5vg0jpN
Code: https://t.co/BzYDlPUpB9
Released 0.1 tensor-sensor via pip, now supports JAX from @froystig, @SingularMattrix, @cdleary, et al at @GoogleAI in addition to @TensorFlow, @PyTorch, @numpy. TensorSensor clarifies exceptions by augmenting messages + visualizing Python code to indicate shape of tensor vars. pic.twitter.com/ckuB9F2bFT
— Terence Parr (@the_antlr_guy) December 3, 2020
If you missed it yesterday, check out the new tutorial on 🤖 supervised contrastive learning 🤖 on https://t.co/m6mT8SrKDD: https://t.co/VxEKjI5uNr
— François Chollet (@fchollet) December 3, 2020
A bunch of researchers in China have built 'CPM', a Chinese language GPT-style model. Paper here: https://t.co/frVcm7TNhC It's notable to not see analysis of bias (when compared to LM papers in the West). Can people point me to good studies of Chinese language bias?
— Jack Clark (@jackclarkSF) December 3, 2020
John Moult says that "Although code-sharing is obviously desirable, and some groups do it, it has never been the norm."
— Lior Pachter (@lpachter) December 2, 2020
Maybe not for CASP contestants but it *has* to be the norm in science as far as I'm concerned.https://t.co/keXz7SIFzO
One more new tidymodels package this week!
— Max Kuhn (@topepos) December 2, 2020
{finetune} lets you tune models using racing methods (fewer model fits) or via simulated annealing.https://t.co/jIul1iJ4Fp#rstats pic.twitter.com/wsme5du34b
Powerful LMs such as GPT-3 and T5 have impressive ability to answer questions by continuing a textual prompt. However, how can we know when an LM knows the answer with confidence, and when it's making a random guess? Our new preprint asks this: https://t.co/Gupb7WP9zS 1/N pic.twitter.com/aahY9P4hZU
— Graham Neubig (@gneubig) December 2, 2020
i cannot tell you how many days & nights we spent with heated discussions on right and wrong ways to analyze what deep nets we built for breast cancer screening do. thanks @taromakino for pulling this throught! he’s one of the first cohort of the medical track at @NYUDataScience. https://t.co/Nq8JrLdM9C
— Kyunghyun Cho (@kchonyc) December 2, 2020
