No Images? Click here
Data Analytics Newsletter #7, June 2019
The Data Analytics Practice Committee (DAPC), the Young Data Analytics Working Group (YDAWG) and the Institute are pleased to bring the latest in the world of Data Analytics to your inbox, and to share some of our recent work with you.
In this edition, we have more fake news, tips on getting started, and articles on combining machine learning with statistical and actuarial approaches.
We hope you enjoy reading the newsletter!
Content Summary
1. Fake News II: Electric Sheep Dreams
- I love the smell of (napalm) French toast in the morning: Editing videos by editing text
- This _______ does not exist
- Interview with Christian Szegedy
2. Mixing It Up: Machine learning with actuarial and statistical techniques
- How Statisticians and Data Scientists could learn from each other
- AI in Actuarial Science: The State of the Art
- Insights - A history of loss reserving leading to granular models and machine learning
3. Analytics Snippet
- Feature Importance and the SHAP approach to machine learning models
4. Getting Started
- 10 Reads for Data Scientists Getting Started with Business Models
- Microsoft: Introduction to R and Python for Data Science
- Matlab / Python / Julia Cheatsheet
- R for Data Science
5. Editor's Note
Fake News II: Electric Sheep Dreams
In the last edition of this Newsletter, we had fake election tweets and writing prompts. We continue the theme with more examples of researches using generative AI models actross images, videos and text:
I love the smell of (napalm) French toast in the morning: Editing videos by editing text
From the world of fake video, Stanford university researchers have developed techniques to make editing video footage as easy as editing text – no reshoots required!
These pictures aren't real, but are impressively realistic nonetheless. The last one may just make you a bit hungry.
Interview with Christian Szegedy
If you’ve got a spare half hour on your commute, check out this Interview with Christian from Google Brain who was the inventor of adversarial training, a key concept behind generative models.
Mixing It Up: Machine learning with actuarial and statistical techniques
Why not both? Articles on combining statistical, actuarial and machine learning techniques.
How Statisticians and Data Scientists could learn from each other
From the Actuaries Summit, Xavier Conort describes how machine learning and traditional statistical techniques can complement each other.
AI in Actuarial Science: The State of the Art
From the South African Colloquium in April, a broad review of machine learning in actuarial science – spanning across applications to both life and general insurance.
The video is also available on Actuview
Insights - A history of loss reserving leading to granular models and machine learning
For those working in general insurance, this presentation from our recent Insights session with Greg Taylor dives further into claims reserving models, and how machine learning can be applied.
Analytics Snippet
Feature Importance and the SHAP approach to machine learning models
Andrew Ngai explores tools to unpack black box models to understand variable significance and explain why a model is making the predictions it does.
Getting Started
10 Reads for Data Scientists Getting Started with Business Models
This article covers a variety of broad topics that may be of interest for data scientists, with a focus on business issues rather than modelling. It makes the point that to be an effective data scientist, it is not enough to have strong technical skills – commercial acumen is critical to success.
Microsoft: Introduction to R and Python for Data Science | R Python
For those new to the data analytics journey, this course by Microsoft in partnership with DataCamp offers a free introduction to coding in R and Python.
Matlab / Python / Julia CheatSheet
For those coming from Matlab in university, this cheatsheet may help the transition to Python or Julia.
An introductory textbook by Hadley Wickham, chief scientist at Rstudio and the brilliant mind behind the dplyr package, and Garrett Grolemund. It covers the data science process, data input and transformation, modelling and options for presenting in R.
Editor’s Note
The best way to learn something is to teach it to someone else. We're continually looking for more writers for Analytics Snippets in Actuaries Digital. Even if you only have writing, R only, Python only skills, we can use your help.
If you have any suggestions, corrections or contributions for the next Newsletter, please do not hesitate to contact Jacky Poon.
Past editions of this Newsletter are now available here.
Disclaimer
The Institute wishes it to be understood that any opinions put forward in this publication are not necessarily those of the Institute.
 |  |  |  |  |
Level 2, 50 Carrington Street
Sydney NSW 2000, Australia
t +61 (0) 2 9239 6100
f +61 (0) 2 9239 6170