Skip to content

DSB #121


hopefully you already went home and now you may read the DSB! I recommend an interesting article about working with databases with SQLAlchemy from Analytical or an article about limits of gradient descent from Computer Science & Science.

And as always, enjoy your reading.

Analytical – This article is the second part from the series that creates a fullstack application with FastAPI, Nuxt and Postgres. In this part you can learn how to work with databases with help from SQLAlchemy. (rcmd by reader) – Optimize
your PySpark solution. – After a long time, a really good list with useful Pandas functions.

Computer Science & Science – Limits of gradient descent algorithm. What makes it struggle and when. – Quite detailed, funny and long description of vulnerability in a fictitious dating app allowing to identify the users exact location. – Find out what your computer science level is in this matrix. Or apply for this online test. (rcmd by reader)

Graphs and Visualizations – Multiple interactive graphs showing how BERT can predict the missing words in the sentence. (rcmd by reader) – I know that all of you are familiar with the coefficient of determination, but this interpretation with help of Venn diagrams is lovely and insightful. – How to visualize ordinal numbers with help of Bayessians methods.

Business and Career – Lol, people investing their money in crypto wild west are surprised that their money can be stolen or lost. – BTM ( Bitcoin ATM) operators are going to require more than just a phone number. – This could be a fatal problem for all sensitive data in the cloud. T-Mobile is facing a lawsuit for its data breach. (rcmd by reader)

Pop – Which companies have the most sensitive data? (rcmd by reader) – Snapchat (yeah, still a thing) released Scan. A feature that identifies things in the real world. Interesting part is that it’s heavily powered by a mix of other companies’ recognition engines. – FB probably wants to create a commision in order to handle elections objectively.

Education – Patterns of confusing explanations used by so many authors in so many articles. Very interesting reading. – Learn about similarity indexes like Flat, LSH, HNSW, and IVF. When to use which of them. – Queues and stacks are Python basics. Learn how to use them effectively.

Data & Libraries – Who’s the owner of the data? Not so simple or is it? (rcmd by reader) – Thanks to the big data we know that graphs are not enough, it’s time for hypergraphs. A higher-order system brings more ways to model. – Simpson’s paradox in practice explaining why the stories about vaccine ineffectiveness are nonsense. (rcmd by reader)

MLOps – What is MLOps and what are the challenges of this new field? – Some sources about MLOps. Well, maybe too many for one person, but it’s just a list.

Video & Podcast – 6 not so long lectures about Deep RL basics. – Recording from userR! conference.

Papers & Books – Large fundation models (BERT, GPT-3) are risky since they can contain many hidden flaws and inject them into other models. The authors of the paper bring some solutions.

Behind the Fence – Senior Data Analyts in HER, USA.


Be First to Comment

Leave a Reply