Skip to content

DSB #106


hopefully it’s Friday and hopefully you’re reading the bulletin! For me the most interesting is a Python library SDV from Data & Libraries that generates fake data. Very useful is also a article about proper „giting“ in Computer Science & Science.

As always, enjoy your reading.

Analytical – Use NLP to build not only text classifiers with language explanations. – Covid is affecting everything, therefore even ML models, so maybe it’s time to retrain them. – We are used to test software, monitor our models, but we should also continuously test our data.

Computer Science & Science – How to git properly? Daniel Stenberg, lead dev of curl, gives you some hints. (rcmd by reader) – Design patterns for software architecture of modern applications – how to meet scalability, availability, security, reliability, and resiliency demands. (rcmd by reader) – The dark side of microservices, which problems they brought and how to solve them. (rcmd by reader)

Graphs and Visualizations – Use Dash for creating classification models with an app.

Business and Career – Roadmap for data engineer. At least it shows you which technical stack is in demand. – In data science, there is new discipline, machine learning engineering that gives answer on question: What can we build with machine learning models developed by data scientist, and how? (rcmd by reader) – Bitcoin is surging (again) and investors are optimistic about it.

Pop – Finally, a reason to have a baby! You can create your own baby monitor. (rcmd by reader) – Covid-19 vaccines are incoming and there should be a strategy for their optimal rollout. – Does Facebook’s AI stop hate speech and misinformations? And here you can read how the moderation works.

Education – What current research tells us about interpretability of ML and what is the future? – Learn about Temporal Difference Learning. – AI is solving partial differential equation and it is 1,000 times faster than traditional mathematical formulas. (rcmd by reader)

Data & Libraries – SDV (Synthetic Data Vault) will help you to generate fake data based on the actual data. – FinRL is a library that helps you to develop your own stock trading strategy with help of deep reinforcement learning. – Alibi Detect is a Python library that detects outliers, adversarials and drifts.

Papers & Books – Summarized Top 10 ML papers in 2020. (rcmd by reader) – Memory optimization for deep networks. – Large neural networks compressing.

Behind the Fence – Data Scientist in The Ohio State University, Columbus, USA.

Joke – Every single time…

Be First to Comment

Leave a Reply