2019 is also finishing with great achievements in AI field. Thanks to the extended report from 'artificial intelligence index' which I highlighted more specific ones (Of cource is just my choise only):
👉 AI Research went crazy. Between 1998 and 2018, there’s been a 300% increase in the publication of peer-reviewed papers on AI.
👉 Attendance at conferences went crazy too, for eg. NeurIPS, got some 13,500 attendees this year, up 800% from 2012.
👉 Education too bumped up, a lot of folks took up MSc / PhD with something in Machine Learning
👉 USA still leads in AI, no matter what other countries say
👉 AI algorithms are becoming cheaper and mainstream
👉 self driving vehicles market is coming of age and raking in a lot of investments

Due to your interest and some specific points, you will read about in more detail from the report:
https://hai.stanford.edu/sites/g/files/sbiybj10986/f/ai_index_2019_report.pdf?fbclid=IwAR228NxD7QCdksNYkSPZ2vcpm5Jzk5zCGx9v0NpsAkQVOspv85MvG3LK3wE

Happy new year.
I would like to share channel's progress for 2019 and we have +29 874 new members for this year. Thank you for all members of channel.

These Projects will make you the Super Hero of Python:

https://medium.com/dataflair/these-projects-will-make-you-the-superhero-of-python-city-14101e62393b

YOLACT (You Only Look At CoefficienTs) - Real-time Instance Segmentation
Results are impressive, above 30 FPS on COCO test-dev

Everybody’s Talkin’: Let Me Talk as You Want

This paper presents a method to edit a target portrait footage by taking a sequence of audio as input to synthesize a photo-realistic video.

Paper pdf:
https://arxiv.org/pdf/2001.05201.pdf

Github:
https://wywu.github.io/projects/EBT/EBT.html

Youtube:
https://youtu.be/tNPuAnvijQk

Synergetic Reconstruction from 2D Pose and 3D Motion for Wide-Space Multi-Person Video Motion Capture in the Wild

Paper:

https://arxiv.org/pdf/2001.05613.pdf

Project page:

http://www.ynl.t.u-tokyo.ac.jp/research/vmocap-syn/

Dataset will be available publicly soon

More than 200 NLP datasets - this is gold (last update 21.01.202)

https://quantumstat.com/dataset/dataset.html

and also Google provided dataset search tool for publicly available datasets:

https://datasetsearch.research.google.com/

End to End Machine Learning: From Data Collection to Deployment.

- Collect and scrape data with Scrapy / Selenium
- Train a deep character CNN for (English) sentiment analysis using PyTorch
- Build an interactive web app with Dash to serve the model in real-time
- Put everything in Docker Compose
- Deploy to AWS on a custom domain name

https://towardsdatascience.com/end-to-end-machine-learning-from-data-collection-to-deployment-ce74f51ca203

If you want to play with the app and see how it looks like:
💻Demo: https://www.reviews.ai2prod.com/

Code:
https://github.com/MarwanDebbiche/post-tuto-deployment

Introducing PyTorch3D: An open-source library for 3D deep learning.
PyTorch3D: Faster, flexible 3D deep learning research

This video is synthetic and was created using deep learning

Course from MIT 6.S191 "Introduction to Deep Learning".
Methods and applications in game play, medicine, language, art, computer vision, robotics and more

ZeRO & DeepSpeed: New system optimizations enable training models with over 100 billion parameters