and surprise, DSB is here! So get comfy and read everything! I would recommend an intro to linear programming from Education, since I personally consider OR-Tools the best package there is for that purpose. Very intriguing is also a sceptical article about personalization from Pop. It’s an unusual opinion to read when some people are able to use crazy words like hyper-personalization.
And as always, enjoy your reading.
https://machinelearningmastery.com/multi-output-regression-models-with-python/ – Article about multioutput regression in Python describing 3 primary strategies how to solve for various outputs. (rcmd by reader)
https://www.kaggle.com/competitions/amex-default-prediction/overview – Kaggle competition with prize money 100 000 USD. Started 11 days ago and the goal is to predict a default of American Express customers.
https://www.andrewheiss.com/blog/2022/05/20/marginalia/ – Comprehensive article about marginal effects and which types there are (AME, MEM, MER…). Contains lots of R code.
Computer Science & Science
https://medium.com/@ari-cohen/how-to-code-a-blockchain-in-6-steps-27fabf5944af – Quick and simple way to code a blockchain. For a more complex solution see DSB #119.
https://fa.bianp.net/blog/2020/polyopt/ – First of five articles about connection between optimization algorithms and polynomials.
Graphs and Visualizations
https://mlu-explain.github.io/ – I love these visual tutorials. The link contains several ML terms and topics explained nicely and understandably with beautiful visualizations.
https://medium.com/@kenanekici/visualizing-multicollinearity-in-python-b5feedc9b3f1 – Visaulization of multicolinearity in Python.
https://www.smashingmagazine.com/2022/05/magical-svg-techniques/ – SVG is a popular format, so it might useful to know how to add texture and depth, how generate SVG pictures, create illustrations or convert raster image into an SVG.
Business and Career
https://blog.pragmaticengineer.com/what-silicon-valley-gets-right-on-software-engineers/ – 7 points why Silicon Valley is much better in utilization of their software engineers than companies in Europe.
https://www.protocol.com/newsletters/pipeline/fintech-startup-valuation-reset – Fintech startups are falling, should we make a wish? Valuation cuts, dismissal of employees…
https://www.fastcompany.com/90751048/in-a-bid-to-step-up-security-and-convenience-google-launches-wallets-and-virtual-cards – Google introduced Google Wallet. It will let you store credentials, payment methods, access tokens (such as digital car keys), and vaccine records.
https://www.marketingweek.com/peter-weinberg-jon-lombardo-personalisation-impersonalisation/ – Sceptical opinion on personalisation, it simply does not work. “Can you name a single famous brand built through personalisation?“
https://erikhoel.substack.com/p/ai-art-isnt-art – In DSB #133 we talked about images generated by DALL-E. This article
argues whether it is art or not. On the other hand some disagree with critique of DALL-E. But no matter the answer, definitely try this hilarious web generating MtG cards.
https://xamat.medium.com/why-did-stars-go-away-a-brief-primer-on-online-user-ratings-6740453f27ed – Is explicit feedback more valuable than implicit feedback? And why has Netflix changed its rating again?
https://mlabonne.github.io/blog/linearoptimization/ – Intro to linear programming in Python with OR-Tools package by Google.
https://www.kaggle.com/code/jhoward/which-image-models-are-best – Intro to timm library which provides state-of-the-art pre-trained computer vision models.
https://www.kaggle.com/code/jhoward/how-random-forests-really-work/ – Notebook from Kaggle about Random Forests on the famous Titanic dataset. Another great and practical intro.
Datasets & Libraries
https://github.com/NannyML/nannyml/#readme – NannyML is a Python library that helps with performance monitoring of deployed ML models.
https://www.philschmid.de/sagemaker-inference-comparison – Learn how to deploy a hugging face transformers with Amazon SageMaker.
https://mymlops.com/ – List of tools you can use for MLOps. Also enables interactively design your own MLOps stack.
https://towardsdatascience.com/what-is-active-metadata-and-why-does-it-matter-add3408c228 – You probably know what is metadata, the next level is active metadata.
Video & Podcast
https://podcastaddict.com/episode/138854855 – What is the difference between MLOps and DevOps?
Papers & Books
https://paperswithcode.com/paper/a-generalist-agent – Gato is another step to generalized AI. It can do multiple things from playing Atari to chatting with you. Is it more than just a big reinforcement learning NN?
Behind the Fence
https://careers-l2tllc.icims.com/jobs/1370/windows-engineer/job – Windows Engineer in L2T, Herndon, USA.