Machine Learning for Everyone.

The best general intro post about Machine Learning, covering everything you need to know not to get overxcited about SkyNet and to get general understanding of all #ML / #AI hype. You can surely save this post into «Saved messages» and forward it to your friends to make them familiar with the subject

Link: https://vas3k.com/blog/machine_learning/

#entrylevel #novice #general

How to read any scientific paper

Reading any paper is hard. And let's admit, it takes time and cognitive resources. Sometimes it may take hours to understand 1 page. If you are constantly skipping or have no clue at all on how to read scientific paper, there is a guide from TNW.

Link: https://thenextweb.com/basics/2019/06/14/how-to-read-a-scientific-research-paper/

#general

​​Using AI to generate recipes from food images

Facebook developed image 2 recipe architecture.

Link: https://ai.facebook.com/blog/inverse-cooking/
Paper: https://research.fb.com/publications/inverse-cooking-recipe-generation-from-food-images/
Code: https://github.com/facebookresearch/inversecooking

#CV #DL

ODS breakfast in Paris! See you this Saturday at 10:30 at Malongo Café, 50 Rue Saint-André des Arts.

You are kindly welcome to submit any news / content you would like to share through the bot, whose message been forwarded above. If you want to increase probability of content being posted, please do follow the structure we use in our channel, providing title, short description and hashtags for the post.

​​A Theory of Fermat Paths for Non-Line-of-Sight Shape Reconstruction

This paper receive best paper award at #CVPR2019.
Main idea: seeing around the corner at non-line-of-sight (NLOS) objects by using Fermat paths, which is a new theory of how NLOS photons follow specific geometric paths.

Link: http://imaging.cs.cmu.edu/fermat_paths/assets/cvpr2019.pdf

​​Accelerating MRI reconstruction via active acquisition

Researchers from #Facebook AI propose a new approach to MRI reconstruction that restores a high fidelity image from partially observed measurements in less time and with fewer errors.

Link: https://ai.facebook.com/blog/accelerating-mri-reconstruction/
Paper link: https://research.fb.com/publications/reducing-uncertainty-in-undersampled-mri-reconstruction-with-active-acquisition/

#CV #DL #CVPR2019 #healthcare #MRI #biolearning

Ranking Items With Star Ratings and How Not To Sort By Average Rating

Two absolute must read articles for proper sorting handling. Sorting items with just an average score is wrong and there is some good classic statistics explanation why.

Link: https://www.evanmiller.org/ranking-items-with-star-ratings.html
Link2: https://www.evanmiller.org/how-not-to-sort-by-average-rating.html

#Statistics #rating #scoring #ranking

Intro to Pythia — Visual Question Answering framework from Facebook

Pythia works in terms of #VQA by taking input picture and question and providing the answer to the latter in terms of picture semantics.


Link: https://link.medium.com/dknDKSuVqX
Previously: https://yangx.top/opendatascience/812

#DL #facebook #pythia #VQA #opensource

Introducing MASS – A pre-training method that outperforms BERT and GPT in sequence to sequence language generation tasks

Researchers from Microsoft Research Asia have introduced MASS—a new pre-training method that claimed to achieve better results than BERT and GPT.

Link: https://www.microsoft.com/en-us/research/blog/introducing-mass-a-pre-training-method-that-outperforms-bert-and-gpt-in-sequence-to-sequence-language-generation-tasks/

#nlp #microsoft

Neural network to undress people on photo

App / web service is released to uncover naked body under the clothes. Works better on swimsuits pictures. Currently website is overloaded and down, updates on the matter are on twitter.

Team's twitter: https://twitter.com/deepnudeapp
Vice article: https://www.vice.com/en_us/article/kzm59x/deepnude-app-creates-fake-nudes-of-any-woman
Website: http://deepnude.com

#HypeNN #DL #CV #nudes

ODS breakfast in Paris! See you this Saturday at 10:30 at Malongo Café, 50 Rue Saint-André des Arts.

Communauté francaise de Machine Learning / Data science - dscientist

Bonjour,
Il existe des communautés anglaises, russes de Machine Learning. En France, il y a de quelques petites communauté (basée sur la localisation ou le domaine spécifique) mais pas une communauté globale. Aujourd'hui dscientist.slack.com est né et permet de regrouper les personnes francophones s'intéressants à ces domaines. L'objectif principal est le partage d'expériences, de connaissances et une veille commune sur l'IA et le ML.

Si vous êtes intéressé(e), rejoignez cette communauté via le lien ci-dessous.

Link: https://docs.google.com/forms/d/1g7v9MpzhgXgDUyAtLDU5MSh78DDaBhGYJgFV1Ys46EU

New free Deep Learning course from #FastAI

This course shows how to build a state of the art deep learning model from scratch.
FastAI’s mission is to make neural nets uncool again by teaching as many people from as many backgrounds as possible.

Link: https://www.fast.ai/2019/06/28/course-p2v3/

#MOOC #DL #freecourse

​​Google announced the updated YouTube-8M dataset

Updated set now includes a subset with verified 5-s segment level labels, along with the 3rd Large-Scale Video Understanding Challenge and Workshop at #ICCV19.

Link: https://ai.googleblog.com/2019/06/announcing-youtube-8m-segments-dataset.html

#Google #YouTube #CV #DL #Video #dataset

​​A #GauGAN timelapse video as a demo for landscape generation

Link: https://www.nvidia.com/en-us/research/ai-playground/

#GAN #Video #Nvidia

ML system used to create complex and accurate simulation of the universe

The speed and accuracy of the project, called the Deep Density Displacement Model, or #D3M for short, wasn't the biggest surprise to the researchers. The real shock was that D3M could accurately simulate how the universe would look if certain parameters were tweaked — such as how much of the cosmos is dark matter — even though the model had never received any training data where those parameters varied.

Link: https://phys.org/news/2019-06-ai-universe-sim-fast-accurateand.html

#Physics #DL #simulation

​​💣New open-source recommender system from Facebook.

Facebook is open-sourcing DLRM — a state-of-the-art deep learning recommendation model to help AI researchers and the systems and hardware community develop new, more efficient ways to work with categorical data.

Link: https://ai.facebook.com/blog/dlrm-an-advanced-open-source-deep-learning-recommendation-model/
Github: https://github.com/facebookresearch/dlrm
ArXiV: https://arxiv.org/abs/1906.03109

#Facebook #DLRM #recommender #DL #PyTorch #Caffe

​​XLNet: Generalized Autoregressive Pretraining for Language Understanding

Researchers at Google Brain and Carnegie Mellon introduce #XLNet, a pre-training algorithm for natural language processing systems. It helps NLP models (in this case, based on Transformer-XL) achieve state-of-the-art results in 18 diverse language-understanding tasks including question answering and sentiment analysis.

Article: https://towardsdatascience.com/what-is-xlnet-and-why-it-outperforms-bert-8d8fce710335
ArXiV: https://arxiv.org/pdf/1906.08237.pdf

#Google #GoogleBrain #CMU #NLP #SOTA #DL

0.2 release of PyTorchPipe

Library for multi-modal deep learning pipelines, in a modular fashion for GPU and CPU.

Link: https://github.com/IBM/pytorchpipe

#IBM #PyTorch