???? Do gay people wait in a LGBT Queue?

Hey folks ??

the economy is taking a nose-dive, so let’s make sure we can do an extra-good job and be un-fireable for the foreseeable future with these awesome papers and libraries!

The Latest Fashion

• I enjoyed this paper about retrieving a meteorite fall with?drones and machine learning!
• Regex is difficult and a bit old. Rulex tries to?fix what wasn’t great?about Regex and transpiles to Regex for maximum compatibility.
• Do you need data for tests, but bundling it in a Github repo is bad? Try [...]

This is "Light to the Party" all links and extra content can be found in the full issue from last week. Want the latest in your inbox? Join 432+ other curious minds.

Hot off the Press

I made a short video about a neat little library that takes matplotlib and makes the API incredibly easy to use, but see for yourself!

Machine Learning Insights

Last week I asked you “How does Amazon recommend new articles to you using machine learning?”, and here’s the answer:

This is a classic recommender system problem. Here, an online retailer assumes that a customer that is similar to other customers will have similar tastes and likes. The problem is that few people have actually bought the same exact products. If we break it down to computer science abstractions, we have a very large sparse matrix with entries for customers and products.

Using those ingredients, we can use an algorithm called “collaborative filtering”. Usually, this is done in two steps:

1. Determine similar users with similar shopping behaviour
2. Estimate which items they’re most likely to recommend

However, this can be quite expensive if the userbase gets large enough and only gives a general “we think you might like this”. Alternatively, we can also build a system in which we recommend what other users also bought. This move from a user-centric to an item-centric approach is as follows:

1. Build a relationship between items that were often bought together by the same customer
2. Match the item-centric matric to the current customer tastes

Naturally, we can build more sophisticated matching approaches that take into account problematic matching, or different customer segments, once we have got this overall approach working.

Question of the Week

• What is target encoding and what do you have to avoid applying it?

Post them on Twitter and Tag me. I'd love to see what you come up with. Then I can include them in the next issue!

Tidbits from the Web

• Evan & Katelyn mad LED dice that light up when they hit the table!
• Google promised to double its AI ethics team, one year later they’re pretty vagues about follow-up.
• Don’t know what to make for lunch? This Tiktok account “Rolls for Sandwich” every day and I’m obsessed.

This is "Light to the Party" all links and extra content can be found in the full issue from last week. Want the latest in your inbox? Join 432+ other curious minds.