Artificial Intelligence #14
Andriy Burkov
PhD in AI, ML at TalentNeuron, author of ?? The Hundred-Page Machine Learning Book and ?? the Machine Learning Engineering book
Hey, this is the fourteenth issue of the newsletter. I would like to thank you for being a dedicated subscriber and, before we start, I need your advice. I want to improve the newsletter and I look for the ideas of simple changes that do not require much time investment. Please share your ideas with me in the comments.
This issue contains a sponsored message from this week's sponsor Lambda Labs. As usually, the sponsored link is labeled with the [Sponsored] label.
- [Tool] PyBoy: an open-source Game Boy emulator written in Python that allows you to train your machine learning model on the game data
- A Chrome extension that adds a 3D photo effect to Instagram pages and based on the 3D-photo-inpainting running in Colab (paper, paper github)
- GANs in computer vision: an article review series on the most popular GAN papers (part 1, part 2)
- [Sponsored] Lambda GPU Cloud — Train Models Faster and Cheaper
Lambda Cloud deep learning instances let you train models 2x faster than a p2.8xlarge for $1.50/hr. Pre-installed with Ubuntu 18.04, TensorFlow, Keras, PyTorch, Caffe 2, CUDA, and cuDNN
- [Video] Efficient PyTorch debugging with PyTorch Lightning
- [Data] The Synthinel-1 dataset: a collection of high-resolution synthetic overhead imagery for building segmentation
- [Tool] Advbox: a toolbox to generate adversarial examples that fool neural networks (paper, github)
- Amazon: "Training a product discovery system on many languages at once improves performance in all of them" (blog article, paper)
- [Report] On machine learning interpretability
- [Book] Linear Algebra Done Right by Sheldon Axler (free to download through July)
- [Article] The economics of AI today
- [Video] Privacy-preserving AI - a talk by Andrew Trask
- [Tool] Spotifyr: R wrapper for pulling track audio features and other information from Spotify’s Web API in bulk
Enjoy the newsletter? Help to make it bigger and better by sharing with your colleagues and friends.
If you want to become a sponsor of one of the upcoming issues, please contact the publisher: True Positive Inc.
Enjoy the weekend! — Andriy
IT Manager Digital Sales Channel - AB Trav och Galopp
4 年Mika Kelja
Python Developer & Data Engineer | Streamlining Processes for Data-Driven Decisions | Converting Business Requirements into Effective Code |
4 年I would like to see a small concluding article on your thoughts would be nice. :)
Automation l Data Science
4 年Idea: Promote an interesting paper that lacks decent code/documentation and encourage subsribers to improve it and share their efforts. Too many awesome papers have bad/out dated code, perhaps this could be a way to improve the situation.