ART: Abstraction Refinement-Guided Training for Provably Correct Neural Networks. arxiv.org/abs/1907.10662

Just tried out popOS at my workplace and got blown away by the smoothness of the whole operating system. This should be the go-to choice of anyone who is looking for a Debian based OS where GPU accelerated deeplearning environments can be set up in no time.
It comes with proprietary Nvidia display driver out of the box

Proprietary drivers can be updated via native update center

It recognized my dual monitor setup out of the box

Installing Nvidia CUDA and CuDNN can be performed by two commands (I know that can be done via anaconda in a single command but I need TensorFlow 2.0)

Installing TensorFlow GPU / Pytorch is a breeze here

The display scaling is far better than ubuntu

So far, subjectively, the overall user experience is smoother than ubuntu

PopOS : https://system76.com/pop
Installing CUDA & CuDNN: https://support.system76.com/articles/cuda/

"Few-Shot Adversarial Learning of Realistic Neural Talking Head Models"
Zakharov et al.: https://arxiv.org/abs/1905.08233
#ArtificialIntelligence #DeepLearning #MachineLearning

Towards Reproducible Research with "PyTorch Hub"

By Team PyTorch : https://pytorch.org/blog/towards-reproducible-research-with-pytorch-hub/

#deeplearning #pytorch #research

The 10 Deep Learning Methods AI Practitioners Need to Apply
https://medium.com/cracking-the-data-science-interview/the-10-deep-learning-methods-ai-practitioners-need-to-apply-885259f402c1

Towards AutoML in the presence of Drift: first results. arxiv.org/abs/1907.10772

Andrew Ng: How to Choose Your First AI Project
https://hbr.org/2019/02/how-to-choose-your-first-ai-project

Beautifully drawn notes on the deeplearning.ai specialization on Coursera, by Tess Ferrandez.

https://www.slideshare.net/TessFerrandez/notes-from-coursera-deep-learning-courses-by-andrew-ng

Google at ACL 2019

https://ai.googleblog.com/2019/07/google-at-acl-2019.html

How Facebook is Using Machine Learning to Map the World Population

https://www.visualcapitalist.com/facebook-machine-learning-world-population-map/

"Neural Point-Based Graphics" - new paper about realistic neural rendering from our lab (Samsung AI Center, Moscow)! We only use pointclouds as proxy for free-viewpoint rendering and do not reconstruct meshes.
youtu.be/7s3BYGok7wU
dmitryulyanov.github.io/neural_point_based_graphics
arxiv.org/abs/1906.08240

Top 8 trends from ICLR 2019

Overview of trends on #ICLR2019:
1. Inclusivity
2. Unsupervised representation learning & transfer learning
3. Retro ML
4. RNN is losing its luster with researchers
5. GANs are still going on strong
6. The lack of biologically inspired deep learning
7. Reinforcement learning is still the most popular topic by submissions
8. Most accepted papers will be quickly forgotten

Link: https://huyenchip.com/2019/05/12/top-8-trends-from-iclr-2019.html

Paper on RoBERTa, the current holder of the pole position on the GLUE leaderboard ( https://gluebenchmark.com/leaderboard/ )

https://arxiv.org/abs/1907.11692

RoBERTa: A Robustly Optimized BERT Pretraining Approach

Liu et al.: https://arxiv.org/abs/1907.11692

#bert #naturallanguageprocessing #unsupervisedlearning

Now AI can also be used to identify fake text.
These researchers have released a tool called has GLTR - Giant Language Model Test Room
1) Test yourself here GLTR tool page: https://gltr.io/
2) GitHub link: https://github.com/HendrikStrobelt/detecting-fake-text
3) Paper link: https://arxiv.org/pdf/1906.04043.pdf

https://t.iss.one/ArtificialIntelligenceArticles

Facebook Research at ACL 2019

https://ai.facebook.com/blog/facebook-research-at-acl-2019/

Deep Learning with MATLAB: Transfer Learning in 10 Lines of MATLAB Code
https://nl.mathworks.com/videos/deep-learning-with-matlab-transfer-learning-in-10-lines-of-matlab-code-1487714838381.html