Hi,
end of the week, start of the weekend and DSB. I can recommend the article about elliptic curve cryptography, really well written, you will learn a lot.
And as always, enjoy your reading.
Analytical
https://fangpenlin.com/posts/2019/10/07/elliptic-curve-cryptography-explained/ – Elliptic Curve Cryptography explained and it’s a really good article, relatively easy to understand with respect to the topic. (rcmd by reader)
https://syncedreview.com/2021/05/14/deepmind-podracer-tpu-based-rl-frameworks-deliver-exceptional-performance-at-low-cost-19/amp/ – Google replaces transformer’s self-attention sublayers with FNet’s Fourier sublayers (simple Fourier transformation), which means 7x faster training on GPU with 92 % accuracy of BERT. (rcmd by reader)
https://session-based-recommenders.fastforwardlabs.com – Session-based recommender based on word2vec.
Computer Science & Science
https://kieranpotts.com/rebranding-rest/ – What is REST [API] and why it should be renamed. (rcmd by reader)
https://ohshitgit.com/ – GIT failures and how to solve them! Hilarious! (rcmd by reader)
https://towardsdatascience.com/javascript-for-data-analysis-2e8e7dbf63a7 – Wow, a serious article about JavaScript for data analysis.
Graphs and Visualizations
http://www.infinitecuriosity.org/vizgp/ – Interactive visualization of Gaussian process. It’s simply amazing. (rcmd by reader)
https://www.wired.com/story/deepfake-maps-mess-sense-world/ – Deepfake maps. Everything will be fake soon.
https://github.com/john-guerra/MIDS_W209_Information_Visualization_Slides – Data visualization course by Berkeley.
Business and Career
https://mcorrell.medium.com/potemkin-data-science-fba2b5ba5cc6 – Is data science really so valuable? If not, you should change it. (rcmd by reader)
https://www.vox.com/recode/22455140/lemonade-insurance-ai-twitter – Lemonade presented itself as a fully AI insurance company and it fires back on twitter where people were alarmed that they will be automatically rejected for their ethnicity, gender and other biases. (rcmd by reader)
https://www.linkedin.com/pulse/nasty-brutish-short-life-modern-cdo-ian-thomas – CDO (Chief Data Officer) is an extremely important role, but unfortunately poorly defined and even now not established or understood.
Pop
https://onezero.medium.com/a-i-is-solving-the-wrong-problem-253b636770cd – AI heavily depends on data and when data is wrong, then AI is wrong. So focus more on quality of data and how and for what you’re using your AI. (rcmd by reader)
https://techcrunch.com/2021/06/02/stack-overflow-acquired-by-prosus-for-a-reported-1-8-billion – StackOverflow is being bought by Prosus. They promis no major changes, so hopefully they’re right since my job depends on it…
https://www.zdnet.com/article/facebook-f8-refresh – Facebook F8 conference is this year focused on developers. Facebook is opening Messenger API for Instagram, updating WhatsApp Business API, launching Facebook Business Suite and it also announced that PyTorch is the default framework for all of its AI.
Education
https://www.freecodecamp.org/news/learn-scikit-learn/amp/ – Scikit-learn course on the freeCodeCamp.org YouTube channel. (rcmd by reader)
https://thenerdstation.medium.com/how-to-become-a-quantum-software-engineer-646ca45c4001 – Where to learn basics about quantum computing.
https://www.pinecone.io/docs/examples/it-threat-detection/ – How to use similarity search to detect rare events like IT threats in cybersecurity.
Data & Libraries
https://desfontain.es/privacy/us-census-reconstruction-attack.html – How does the US Census protect its data from reconstruction attack? Badly. According to the article anyone can reconstruct Census records. (rcmd by reader)
https://emilyriederer.netlify.app/post/data-error-gen/ – Data generating process (DGP), what is it for and its risks for data quality.
https://www.analyticsvidhya.com/blog/2021/05/greykite-time-series-forecasting-in-python/ – How to use GreyKite, a new library for time series forecasting. (rcmd by reader)
MLOps
https://towardsdatascience.com/lessons-on-ml-platforms-from-netflix-doordash-spotify-and-more-f455400115c7 – When you’re building a ML platform you need to decide what to buy and what to build by yourself. It’s interesting how many tech firms have their own in-house solutions and tools. (rcmd by reader)
https://veselinastaneva.medium.com/what-ive-learned-about-mlops-from-speaking-with-100-ml-practitioners-3025e33458ad – Takeaways from interviews with more than 100 ML practitioners about mlops and why more than 80 % models never see the production.
https://www.morningbrew.com/emerging-tech/stories/2021/05/26/regulate-ai-just-play-sandbox – AI sandbox, a place where safe and privacy-friendly AI can be built. Sandbox will be a platform to test an algorithm for bias without getting the company in trouble.
Video & Podcast
https://www.youtube.com/watch?v=Wu0hAjlMqUQ&t=2572s – Problem of null hypothesis testings and why one should never believe results from only single paper.
Behind the Fence
https://warnermediacareers.com/global/en/job/184464BR/Senior-Data-Scientist – Senior Data Scientist in WarnerMedia, New York, USA.
Be First to Comment