Issue #196
March 27, 2023
Dear friends,
Welcome to the March 26th edition of the Sunday Briefing. This week we’re on hiatus from blogging. While we work on our next post, you can catch up on the latest post in the Viz4Sci series: Waterfall Chart. The latest Medium post “Top 10 Books we read in 2022” and the most recent post on the G4Sci series: Network Attacks: Breaking up a network without observing it completely.
Several of you have reached out in recent weeks asking how they can help support the work we do here at Data For Science. While the Sunday Briefing is a labor of love and will always remain free, it is not without costs, and any help is appreciated.
If you wish to support our work, there are several options:
- Paid subscription to Graphs for Science or Visualization for Science which gives you complete access to the archive of previous posts.
- Subscribe to Medium using our referral link at no extra cost to you
- Directly through a one-time PayPal donation.
Don’t forget we have the very first edition of the Interactive Data Visualization with Python webinar coming up in less than 24h, but there’s still time to Register so you don’t miss out on learning everything you need to know about generating interactive visualizations for your data using matplotlib, bokeh and plotly.
On our regularly scheduled content, we learn how to build a Semantic Search Engine With OpenAI and Pinecone, how to run 100B+ language models at home, BitTorrent‑style and explore a Beginner’s Guide to Synthetic Data.
On the academic front, we dive into strong connectivity in real directed networks, a manifesto for applying behavioural science and the Correlated Impact Dynamics in Science.
This week’s Data Science Book is “The Recursive Book of Recursion” by Al Sweigart. As always you can find all the previous book recommendations on our website. In the video of the week, we have a tutorial that teaches you How to build GPT: from scratch, in code, spelled out.
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. This will help us spread the word!
Semper discentes,
The D4S Team
This week’s Data Science Book is " The Recursive Book of Recursion " by Al Sweigart, a highly recommended book for programmers of all levels. The book explains recursion in a clear and approachable way. It begins by laying important groundwork and explains functions and their operation and features. The author spends significant time explaining the call stack, what it does, how it is structured, and how it operates, leading to a discussion of ‘stack overflow,’ one of the risks of using recursion. He then devotes an entire chapter to comparing recursion and iteration, demonstrating that in the vast majority of cases, recursive functions are not necessary and in some cases perform worse than their iterative counterparts.
However, the book also shows where recursion is actually a good idea and where it is a good fit. Sweigart explores traversing tree structures and demonstrates how memoization can improve the efficiency of some recursive functions. The book ends with several projects that build on the concepts that come before, including the Droste Effect, a recursive art technique that generates a similar recursive image from any photograph or drawing utilizing images.
Overall, the Recursive Book of Recursion is a great read for beginners and intermediate programmers alike. The book teaches about recursion and stretches the reader to think differently while confidently showing that seemingly lofty concepts are within reach.
- 1. Building a Semantic Search Engine With OpenAI and Pinecone [sigmoidprime.com]
- 2. Run 100B+ language models at home, BitTorrent‑style [petals.ml]
- 3. The genie escapes: Stanford copies the ChatGPT AI for less than $600 [newatlas.com]
- 4. The Unpredictable Abilities Emerging From Large AI Models [quantamagazine.org]
- 5. Leveraging Time-Series Segmentation and Machine Learning for Better Forecasting Accuracy [odsc.medium.com]
- 6. A Beginner's Guide to Synthetic Data [pub.towardsai.net]
- 7. Recipe for Disaster: The Formula That Killed Wall Street [wired.com]
- • Strong connectivity in real directed networks (N. Rodgers, P. Tiňo, S. Johnson)
- • Spatial immunization to abate disease spreading in transportation hubs (M. Mazzoli, R. Gallotti, F. Privitera, P. Colet, J. J. Ramasco)
- • A manifesto for applying behavioural science (M. Hallsworth)
- • Superhuman artificial intelligence can improve human decision-making by increasing novelty (M. Shin, J. Kim, B. van Opheusden, T. L. Griffiths)
- • Linking social network structure and function to social preferences (J. B. Brask, A. Koher, D. P. Croft, S. Lehmann)
- • Correlated Impact Dynamics in Science (J. Liu, T. Kunal, D. Wang, C. Song)
- • From localized to well-mixed: How commuter interactions shape disease spread (A. Winn, A. Konkol, E. Katifori)
Let's build GPT: from scratch, in code, spelled out
All our videos are also available in our
YouTube playlist.
Enjoy the newsletter?
Forward it to a friend, or subscribe to get it straight to your inbox.
Subscribe Free
