ISSUE 405  ·   September 20, 2022

Trends

Generative AI: A Creative New World

A new class of large language models is making it possible for machines to write, code, draw and create with credible and sometimes superhuman results. It's all part of an emerging field called "Generative AI" and it's blowing up fast. With the help of a GPT-3 author, this article explores the field's rapid growth, the issues, and what to expect next.
Sequoia | Sonya Huang, Pat Grady and GPT-3

Sponsored Link

Dive into the latest in search and ML with luminaries Daniel Tunkelang and Grant Ingersoll 

Join engineers and data scientists from around the world in this 4-week course on search with ML. Learn best practices in: ranking, content and query understanding, and vector search. The course is applied and project-based, with loads of interaction amongst the learning community, and live time with Daniel (fmr head of search at LinkedIn) and Grant (fmr CTO of Wikimedia Foundation and Lucidworks). Class starts in Oct!

Tutorials, Projects & Opinions

Linear Regression

Next up in the popular MLU-Explain series, Jared Wilder offers an awesome introduction to (almost) everything you should know about linear regression. This is a great explainer with interactive visuals.
MLU-Explain | Jared Wilber

 A Sequel to SQL? An introduction to Malloy

Malloy is a new query language that's being developed by Lloyd Tabb, founder and former CTO of Looker. Malloy addresses some of the aesthetic concerns that plague SQL but more interesting is its integration of a query language and a semantic layer into a single language. In this overview, Carlin Eng explores Malloy and how its semantic layer could radically change analytics.
Carlin Eng

Correlation Can Ruin Interpretability

Correlated features can ruin interpretation and it's a more common problem than you might think. In his latest post, Christoph Molnar explores how to identify if an ML model interpretation is affected by correlation, why that's a problem, and how to deal with it.
Mindful Modeler | Christoph Molnar

Operationalizing Machine Learning

Great read for anyone trying to do ML in production. Following interviews with 18 machine learning engineers working on a variety of applications, this paper summarizes common practices for successful experimentation, deployment, and sustaining production performance. Includes pain points and anti-patterns.
arXiv | Shreya Shankar, et al.

Code & Tools

Daft - Python DataFrame for Complex Data

Daft enables you to run both large analytical queries and powerful Complex Data algorithms from the same interface via a Dataframe API.
GitHub | Eventual-Inc 

The Jupyter+git problem is now solved

Previously, using git with Jupyter could create conflicts and break notebooks. With nbdev2, that problem has been totally solved.
fast.ai | Jeremy Howard

 ipyvizzu-story - Animated Charts for Jupyter

With ipyvizzu-story, you can build, present and share interactive presentations in any computational notebook using Python.
GitHub | vizzuhq 

Career

Deciding if a data leadership role is something you actually want to do

If you've been looking for a career change and eyeing a "Head Of" role, this is a great post to help you think through a decision. It's good to get out of your comfort zone but these roles can be very different from one organization to the next. Make sure you know what you're getting into!
Katie Bauer

How did you like this issue of Data Elixir?

👎    1     2     3     4     5   👍

Was this email forwarded to you?  Sign up here >>