[ML UTD 6] MACHINE LEARNING UP-TO-DATE

stitch fix explaining its data science business structure

Here’s Machine Learning Up-To-Date (ML UTD) 6 from the LifeWithData blog! We help you separate the signal from the noise in today’s hectic front lines of software engineering and machine learning.

LifeWithData aims to consistently deliver curated machine learning newsletters that point the reader to key developments without massive amounts of backstory for each. This enables frequent, concise updates across the industry without overloading readers with information.

ML UTD 6 brings updates in the areas of publications, computing, business, and visualization.


ML UTD 6


[Computing] An Image API for…The Smithsonian!

The Smithsonian API documentation
Apparently the government does public APIs now [source]

The Smithsonian Museum (there’s actually eleven) has released an open-access API for metadata on all of its items. The repository is updated on a weekly basis, which is a pretty nice commitment for a largely non-technical organization. Even better, they have great documentation for it as well!


[Computing] Sentiment Analysis on iOS

Screenshots of an iOS app that performs sentiment analysis
View of Omar’s app [source]

Omar M’Haimdat made a writeup on Medium about a cool Sentiment Analysis iOS app he made. The writeup also includes good backstory information on transformers, NLP, and the Hugging Face. Check out his Github repository and an accompanying Colab notebook for training the model.

Aside from reading this article, I also highly recommend checking out the Hugging Face transformers library. It has state of the art Tensorflow and PyTorch NLP models ready to go.


[Computing] DeepMind Releases JAX-Based Libraries

A screenshot of the documentation for the Haiku software library from DeepMind
DeepMind was nice and made good docs [source]

You’ve probably been hearing more and more about the JAX library for neural network primitives (and for good reason). You’ll be happy to hear, then, that DeepMind has released some new libraries for building on top of JAX. One is Haiku, which the team aptly brands as “Sonnet for JAX”. This library aims to be a general-purpose solution for building neural networks from the JAX primitives. The other library is RLax, pronounced “relax”. As you can probably guess, this library caters specifically to training reinforcement learning agents.


[Business] StichFix Explains its Innovative Structure

A screenshot of an article by Stitch Fix detailing their data science business structure
Stitch Fix takes a brand new approach on Data Science in business [source]

I’ve always thought Stich Fix has a cool product, delivering personalized clothing choices that just make sense. I now love them even more, after reading their article about how they structure the business. The article details how they truly treat data science as a first-class citizen. I mean this literally, too — the Data Science department is equated with Marketing, Sales, etc. They have a Chief Algorithms Officer. This is an awesome effort, and I want to see more examples of companies trying similar things. Let me know if you find any!


[Business] Cibo Express Opens First Auto Checkout

A picture of CIBO Express Gourmet Market
No cashier, no problem! [PRNewsfoto/OTG]

Flying into Newark Liberty Airport anytime soon? If so, head over to the Cibo Express in Terminal C. You’ll find that the store is operating the first retail implementation of Amazon’s new no-checkout purchase experience. This comes just days after Amazon announced the retailing of its disruptive service.


[Visualization] Machine Learning for Everyone

A comic about a polarity of how people perceive machine learning
Well ain’t that the truth [source]

Machine Learning theory can get complicated; there’s no two ways about it. Every once in a while, an article comes along that presents some of the details so clearly, and it feels so good. This is one of them. Share this guy off to those struggling to become proficient, to help them fill the gaps with intuitive explanations. Then, share it to everyone else anyways, since it’s entertaining and a nice review.


[Visualization] ML Weather Prediction

Precipitation drives taxi demand shifts — visualized.

After finding the article above, I checked out this guy’s whole website. Here‘s another great article! This one explains the intricacies of weather prediction. Notice how taxi traffic (the colors) shifts with precipitation. How’s that for a business model? Send this to your Uber-driving friends and see if they agree!


Stay Up To Date

That’s all for ML UTD 6. However, things happen quickly in academics and industry! Aside from ML UTD, keep yourself updated with the blog at LifeWithData.

If you’re not a fan of email newsletters, but still want to stay in the loop, consider adding lifewithdata.org/blog and lifewithdata.org/tag/ml-utd to a Feedly aggregation setup.

0 Replies to “[ML UTD 6] MACHINE LEARNING UP-TO-DATE”

This site uses Akismet to reduce spam. Learn how your comment data is processed.