Skip to content

DSB #128


the first volume of this year comes on Sunday evening, and brings lot’s of reading. I found the article about logging from Computer Science & Science very practical and maybe more generally I would recommend focusing on web 3.0 which is discussed almost everywhere and also in our Pop category.

And as always, enjoy your reading.

Analytical prepares movie credits. And they order them with the help of graph theory. (rcmd by reader) – Generalization hase been an important topic in ML for years. How do we define it for language models? – Quite nice and simple exploratory graph analysis of 6 milion NFT transactions with code. (rcmd by reader)

Computer Science & Science – To log, or not to log? And what are the alternatives when ordinary logging is unsuitable? (rcmd by reader) – Recoginze good error messages and learn some best practices on how to write them. – Did you know that numbers in Python have methods or a hieararchy? And that’s not all…

Graphs and Visualizations – Investigate the papers published at NeurIPS in the last 35+ years. Full of inspirational ideas on visualizations. – Wow, not only links, but also description of 100 beautiful notebooks of last year. Another inspirational article. – It takes skill to create a great graph, in this twitter thread you can read and see some differences between an average graph and an amazing graph.

Business and Career – Scaling in Agile, how to handle complexity of the whole process and when you should just avoid it. “Agile is primarily about empowering those responsible for creating value... so that they can deal with the scaling processes and tools themselves.” (rcmd by reader) – Can you cluster and distinguish companies by data to engineers ratio? Yes, you can and you will get an interesting story. – U.S. banks plan to offer their own stablecoin, called USDF. Question is whether it will be insured or not.

Pop – Is Web 3 a real thing? And what can we expect from this so-called revolution? – Top 10 articles about AI from last year. Some of them indicate sobering from uncritical hype. (rcmd by reader) – After 20 years an
Italian mafia boss was arested thanks to Google Street view. (rcmd by reader)

Education – Build your own neural network from scratch! I like these hands-on tutorials. – Intro on
adversarial attacks and defenses. (rcmd by reader) – Exhausting intro to variational autoencoders, mainly theoretical with lot’s of equations.

Data & Libraries – Change your data warehouse into centralized operational brain with entity layers and reverse ETL. – Dapr = The Distributed Application Runtime. It provides APIs that handle microservice connectivity. (rcmd by reader) – VisiData is an open-source tool for datasets handling in terminal. Demo video is here. (rcmd by reader)

MLOps – Data Engineers at Google treat SQL the same way Software Engineers treat code. Which of course does not mean only obvious tools like Git, but mainly standardized approaches. – See how the system designs for industrial recommendations and search works. Examples from companies like Alibaba, Facebook and more. (rcmd by reader) – Design of real-time ML. What can go wrong and how to handle it.

Video & Podcast – The 2021 AI rewind in 15 minutes.

Papers & Books – Deep Learning Interviews is a book, or rather an inventory of numerous job interviews and exams. – 50 reaserch papers, 40 blog posts, 11 new articles, 9 webinars, 8 events and 3 podcasts just about soccer analytics. – The top trending papers, libraries and datasets for 2021 on Papers with Code.

Behind the Fence – Senior Analytics Engineer at Fabulous in Paris, France.


Be First to Comment

Leave a Reply