Data-Centric AI for Ranking
Something that I’ve heard anecdotally is that data scientists often jump right into model optimization before addressing the data. In reality, it should be the entirely opposite order of operations. The article encourages building a strong domain knowledge through collecting, cleaning, and feature selecting prior to doing anything with a model. - John Forstmeier
Decentralized Data Engineering
The author argues that with a focus on automation and templates, data teams can become more decentralized by providing interoperability between tools, quicker delivery with self-service data platforms, and increased standardization. - Blake Burch
The statistical magic behind the bootstrap
Came for the Bootstrap, stayed for the math. I wanted to see how similar the name is to software bootstrapping, and turns out it’s the same. From a mathematical perspective, the article provides good insight into a quasi proof of concept approach for determining a statistic without nailing down the real theory. - Eric Elsken
Data Skills Can Make a Big Difference in Non-Data-Science Careers
This article is an interesting dive into how the study of data science provides skills that are transferable to other fields. One thing that Himalaya Bir Shrestha, the interviewee, talks about is how “data science lies at the intersection of mathematics, computer science, and domain expertise” and how that expertise can be in just about anything. - Katt Baum
AI: The Tool, Not the Movie
Just a fun and interesting read on why the way AI actually works and is used, is totally different than the way AI is portrayed in movies. - Jon Davidson
Image Recognition: Past, Present, and Future
As Dall-E has gained popularity on social media, I think it is interesting to look at the history of image recognition. This article provides a great brief history on image recognition. - Steven Johnson