Well this is impressive: Translating pose from person to person in video arxiv.org/abs/1808.07371 youtube.com/watch?v=PCBTZh…

Singular Values Of Convolutional Layers arxiv.org/abs/1805.10408 Maybe time to get a new twitter photo with a better method :) Back in 2014 I was playing with power iterations to obtain the spectrum of the first layer of OverFeat. My Twitter photo is its singular vectors.

Yes! Thank You! Exporting the beauty of numpy syntax into the world!

somebody on twitter (can't find) once said that setting up travis ci is like text chatting with a monkey on a space ship ... -- definitely true -- ... I sped up my side of the comm by epsilon using the github edit function on a throw-away branch until it worked 🙉

Over the last couple of months some friends and I made Kymatio, the Python package that brings Wavelet Scattering into your deep learning pipelines. It can handle 1D, 2D, and 3D signals. Give it a spin! x.com/kymatiowavelet…

Dear Reviewers, our rebuttal didn't fit in 5000 characters. Run this: echo H4sIANqgilwAAx3Muw7CMAxG4T1P8Y90CSuXjYmRtaMBq4kIseQ4rfL2mK7nk84sHcpUyoBShaXcImaPnC2xhgNN+FTZsCUyDAdSxltyXa4Q9+e0l5V1wLQ3c4kId/aYmw8Z9LJO5f+uyyW0JGpdSyQ7fm+P0zn8AAUwpEaEAAAA | base64 --decode | gunzip

We managed to fit a >10000 char rebuttal into a single tweet x.com/meickenberg/st…

2. Matthew Rocklin: The Python community fractures when we develop competing packages that are not compatible with each other and thrives when we develop common APIs. Try to follow already existing APIs when you can. (Also GPUs)

Starting fall 2019, I will join the Department of Biomedical Engineering at McGill's Faculty of Medicine. Please mail me ([email protected]) for PhD, post-doc, or research engineer positions in my new group in Montreal. Neuroskeptic 🇺🇦 @neuroconscience The Neuro

Colaboratory I still don't quite understand why @googlecolab develop their own front-end to Jupyter, which is still behind #JupyterLab in terms of features at the moment.

To continue the thread:I pledge to continue opposing racism&hate, including systemic racism in science. And to do this work everyday. Join me by nominating 5 others to do the same ✊🏿✊🏾✊🏽cc Valentina Borghesani Fabian Pedregosa Pedro Pinheiro-Chagas Matteo Visconti di Oleggio Castello Sara Popham x.com/alex_ander/sta…

The NumPy paper is out! nature.com/articles/s4158…

nature.com/articles/s4158… About time! Congrats to the authors!

Vlad is not only a top researcher, but also one of the kindest persons I know. Can't think of a better advisor.

New paper about links between #DeepLearning and the brain's visual system by Michael Eickenberg Parietal, INRIA: sciencedirect.com/science/articl…

Stochastic Subsampling for Factorizing Huge Matrices Our preprint with full proofs & empirical study (10x speedups) hal.archives-ouvertes.fr/hal-01431618v1…

Torch for python! github.com/pytorch/pytorch

Very excited to have been awarded a Marie-Curie Global fellowship. #MSCA I'll spend two years at UC Berkeley and third one at ETH Zurich

Question about permutations and permutation testing: Is it OK or even desired to use only derangements (en.wikipedia.org/wiki/Derangeme…) in a permutation test? If so, why? If not, why not?