Issue #54

June 7, 2020

Dear friends,

Welcome to the 1 year anniversary edition of the Sunday Briefing.

The first ever issue of this newsletter went out on June 2nd, 2019 . It’s been a long year since then, with many changes and improvements along the way. When the first issue went out, I could count the number of subscribers in the fingers of my hand, and now we’re several thousand strong.

The first two issues were actually called “Data for Science Newsletter” as we didn’t official name it “Sunday Briefing” until Issue #3 on June 16. The format of the newsletter has also slowly evolved over time. The video of the week section didn’t start until Issue #13 on Aug 29, 2019. This top blurb didn’t include links until Issue #32 on Jan 5, 2020 and the blog didn’t really get going until Issue #36 on Feb 2, 2020. Finally, the CoVID blog series started on Issue #44 on March 29, 2020. Along the way we managed to (somehow) never miss a week, and learn many new ideas and concepts. There were many successes (and some failures) and we’re looking forward to what this second year will bring.

This week we are taking a break from blogging, but you can still check out our latest blog post in the CoVID-19 series: Visualizing the spread of CoVID-19 In this post we take a deep dive into the Johns Hopkins University data repository and some of the visualizations and analyses that can be made with it. As always, you can follow along with the GitHub repository containing the respective Python code. We hope you find it useful and gladly welcome any comments you might have.

In our regularly scheduled programming, we take a beginners mind and dive into the fundamentals with Practical Python Programming , and guides to Python Debugging and Deploying Machine Learning Models . We’re also happy to introduce you to Acme: A new framework for distributed reinforcement learning and Altair, a declarative statistical visualization library for Python .

On the academic front, we focus on Natural Language Processing and Computational Linguistics with Language Models are Few-Shot Learners , Emergent linguistic structure in artificial neural networks trained by self-supervision . We also have a A primer on A rtificial neural networks for neuroscientists and machine learning approach for forecasting hierarchical time series .

Finally, in the video of the week, Laurence Moroney and Karmel Allison take us from Zero to Hero in Machine Learning in their tutorial from Google I/O 2019.

Data shows that the best way for a newsletter to grow is by word of mouth, so if you think one of your friends or colleagues would enjoy this newsletter, just go ahead and forward this email to them and help us spread the word!

Today, more than ever,

Semper discentes,

The D4S team

1. Meta-Graph: Few-Shot Link Prediction Using Meta-Learning [eng.uber.com]
2. Exciting Applications of Graph Neural Networks [blog.fastforwardlabs.com]
3. Altair is a declarative statistical visualization library for Python [github.com/altair-viz]
4. Ultimate Guide to Python Debugging [martinheinz.dev]
5. Understanding Programs Using Graphs [engineering.shopify.com]
6. Understanding epidemiology models [arstechnica.com]
7. The Ultimate Guide to Deploying Machine Learning Models [mlinproduction.com]
8. Practical Python Programming [dabeaz-course.github.io]
9. Acme: A new framework for distributed reinforcement learning [deepmind.com]

• Emergent linguistic structure in artificial neural networks trained by self-supervision (C. D. Manning, K. Clark, J. Hewitt, U. Khandelwal, O. Levy)
• Mapping the co-evolution of artificial intelligence, robotics, and the internet of things over 20 years (1998-2017) (K. Börner, O. Scrivner, L. E. Cross, M. Gallant, S. Ma, A. S. Martin, E. Record, H. Yang, J. M. Dilger)
• Agreement threshold on Axelrod's model of cultural dissemination (P. MacCarron ,P. J. Maher, S. Fennell, K. Burke, J. P. Gleeson, K. Durrheim, M. Quayle)
• Revealing Dynamics, Communities, and Criticality from Data (D. Eroglu, M. Tanzi, S. van Strien, T. Pereira)
• Language Models are Few-Shot Learners (T. B. Brown, B. Mann, N. Ryder, M. Subbiah, J. Kaplan, P. Dhariwal, A. Neelakantan, et al)
• A machine learning approach for forecasting hierarchical time series (P. Mancuso, V. Piccialli, A. M. Sudoso)
• Explainable Artificial Intelligence: a Systematic Review (G. Vilone, L. Longo)
• Artificial neural networks for neuroscientists: A primer (G. R. Yang, X.-J. Wang)

Machine Learning Zero to Hero (Google I/O'19)

June 7, 2020

​