AI Research Topics from 1955 https://t.co/wbOTamXyhe https://t.co/y6T2bTebOp
— hardmaru (@hardmaru) August 1, 2018
AI Research Topics from 1955 https://t.co/wbOTamXyhe https://t.co/y6T2bTebOp
— hardmaru (@hardmaru) August 1, 2018
Please don't read this article describing how to beat me in Kaggle competitions. https://t.co/c2kOx97E6C
— Jeremy Howard (@jeremyphoward) August 1, 2018
One mistake I notice data scientists make in writing reports/blog posts is giving "reverse clickbait" titles
— David Robinson (@drob) August 1, 2018
Good: Brad Pitt and Angelina Jolie Get Married
Clickbait: You Won't BELIEVE Which Two Celebrities Just Got Married
Reverse Clickbait: Update on Pitt/Jolie Domestic Status
Quilted Darth Vader, by Xerox PARC researcher and https://t.co/ktYtgBpxGr alum @KalaiRamea Nice interview about her work with computational creativity:https://t.co/Wh3DOGN6ST pic.twitter.com/2Yv932saHA
— Rachel Thomas (@math_rachel) July 30, 2018
Heh - funny how much things have changed since May 2017! https://t.co/aajuiSKZik pic.twitter.com/nAyM8kXDcX
— Jeremy Howard (@jeremyphoward) July 27, 2018
Good luck with NIPS rebuttals! Some discussion and tips:https://t.co/oaj8I0uQWz https://t.co/5K2Yoon6c0 pic.twitter.com/60yh7OoI4K
— hardmaru (@hardmaru) July 27, 2018
This ACLU story is an example of the type of 'empirical investigation' that I think many news organizations should adopt when analyzing aspects of AI. Don't listen to us (the companies), just use our tech and test it. https://t.co/5Iix7vk3jn
— Jack Clark (@jackclarkSF) July 26, 2018
Deep learning and free software: Interesting article about whether neural net weights are considered free (like the GNU-GPL licenses). I think people have not considered neural nets (“Software 2.0”) back when open source licenses were originally developed. https://t.co/nW489Dcqur
— hardmaru (@hardmaru) July 26, 2018
Excellent post by @elenatej on the data science team structure @Airbnb. They found splitting between analytics, inference, and algorithms helped:
— Emily Robinson (@robinson_es) July 24, 2018
1) clarify roles to business partners
2) make analytics people feel valued
3) tailor performance evaluations https://t.co/V7x8ZkkP69
Alphabet Of The Obsolete#innovation #future #EmergingTech @gvalan @sebbourguignon @schmarzo @HaroldSinnott @evankirstel @KirkDBorne @IrmaRaste @Ronald_vanLoon @dhinchcliffe @JGrobicki @diioannid @ahier @MarshaCollier @DrJDrooghaag pic.twitter.com/YVXKaFaF0Z
— Dr. Kash Sirinanda (@kashthefuturist) July 24, 2018
Anyone interested in #AutoML, especially as applied to #DeepLearning, should read this new post by @math_rachel from @fastdotai. Rachel consistently delivers thoughtful, no-BS perspectives on many topics in #MachineLearning and #DeepLearning. 👏https://t.co/NULD6qInsb
— Erin LeDell (@ledell) July 24, 2018
The idea of classifying countries as having high birth rates/high infant mortality rates vs. low birth rates/low infant mortality rates is completely outdated.
— Rachel Thomas (@math_rachel) July 23, 2018
Left chart is from 1965, right is 2017
(from Hans Rosling's Factfulness, arguing against labels developing/developed) pic.twitter.com/f68ssQdcSB