Hello all! It's time for a new edition of my Data Science newsletter!

When it comes to evaluation metrics, the concepts of sensitivity and specificity are always a bit complicated to grasp properly. Perhaps this visualization with distributions will provide you with a new perspective on the subject.

We continue with a news article from the BBC detailing a court ruling made a few weeks ago on the ability for an AI to file patents 🤖. Verdict in the UK and the US: it's a no-go! However, other countries (Australia, South Africa) have provided a positive answer to the question.

For the main course, I suggest you dive into a scientific article from Google Research, both fascinating and very accessible, which presents the concept of "data cascades", where data quality issues have a significant negative impact on the performance and usefulness of machine learning models. Don't be intimidated by the number of pages, it reads very well!

I wish you good reading and a good week! 🙂