The Japanese concept of #ikigai boils down the search for meaning in life into four questions.
— Simon Kuestenmacher (@simongerman600) March 3, 2019
1) What do you love?
2) What are you good at?
3) What does the world need from you?
4) What can you get paid for?
I'm not too far off at the moment...
Source: https://t.co/KREfDKxN2P pic.twitter.com/8tgFHh4hn4
Apparently some people think it's a good idea to listen to me.
— Jeremy Howard (@jeremyphoward) March 3, 2019
Whilst I can't support that proposal, if you do for some reason think it's a good idea, there's now a list of the things I say to do in lessons 1-7 of @fastdotaihttps://t.co/e2UFi9Fv9U pic.twitter.com/Gjat9dQU55
When I was applying to colleges my dad (immigrant from Zimbabwe) told me to not bother applying to Harvard because it was "only for geniuses". He (and I) didn't realize getting accepted wasn't about raw intelligence but playing the application game. https://t.co/joEWBLnGAM
— Chris Albon (@chrisalbon) March 3, 2019
Is Ethical A.I. Even Possible? By @CadeMetz extreme title for a thought-provoking article. Best line: “the foxes are guarding the hen house.” https://t.co/izPn8f5f2k
— Oren Etzioni (@etzioni) March 3, 2019
Kudos to whoever on the @TensorFlow team wrote their Adam optimizer docs :) Thanks for mentioning `eps`, that much under-appreciated hyperparameter!
— Jeremy Howard (@jeremyphoward) March 3, 2019
(This is a slide from a talk I'm doing at Berkeley tomorrow about training neural nets.) pic.twitter.com/jTCjNAcYwQ
There’s no perfect solution to this problem, but I think the tidyverse does a pretty great job
— David Robinson (@drob) March 1, 2019
Strategy as I see it is:
-each function does one thing well
-organize into packages centered around one area
-organize into an ecosystem we call “tidyverse” https://t.co/bMme1sxxV9
As an analytics team, you need to be wary of just ending up doing reactive work, answering quick, tactical “what” questions. Ask “what decision are you trying to make?” and learn to say no if won’t make a difference. Take time to do proactive work.
— Emily Robinson (@robinson_es) March 1, 2019
- @geraldvdb pic.twitter.com/LBoFAHWgn1
Core principles of statistics. what is missing?
— karl rohe (@karlrohe) February 28, 2019
1) sample isn't population. model isn't estimator
2) bias-variance trade off governs all estimation
3) so much *cannot* be made math-rigorous, but we have fun trying
4) randomization is kind of like religion. both can save you https://t.co/ED3ivkY6Qn
Why I avoid Slack, and encourage developers to hold technical conversations on GitHub. https://t.co/yxCVgzhnjU
— Matthew Rocklin (@mrocklin) February 28, 2019
Mastering Fast Gradient Boosting on Google Colaboratory with free GPU
— ML Review (@ml_review) February 28, 2019
By @besab @CatBoostML
Many think GBDT cannot be efficiently accelerated by a GPU, but this is actually not true.
Yandex CatBoost runs 4x faster on Tesla K80 vs Intel Xeon E5–2600v3 https://t.co/TJ5gg9tkld pic.twitter.com/PTEdlMBnHn
One commonly cited argument about the difficulty of learning common-sense reasoning is that "no-one writes down common sense". A counter-argument is "well, the web is big": https://t.co/qPNmra86ES pic.twitter.com/2c721qlTlW
— Graham Neubig (@gneubig) February 27, 2019
The best fistbump moment in academia is when you email a colleague to sheepishly ask for an extension on an agreed deadline...and they're relieved cos they're way behind too
— Chris Chambers (@chrisdc77) February 26, 2019
It's a poignant reminder that everyone is overworked and under the pump basically all the time
