Nice -- GCP's AI Platform Notebooks now support R https://t.co/9nxdNMH7mD #rstats
— David Smith (@revodavid) June 26, 2019
Nice -- GCP's AI Platform Notebooks now support R https://t.co/9nxdNMH7mD #rstats
— David Smith (@revodavid) June 26, 2019
recruiter: Hi, I work for [company] and would like to talk to you about a role on [small "elite" team that rejected you with extreme prejudice a couple of years ago] led by [semi-prestigious person who voted no-hire]. Can you send me your phone number?
— Joel Grus ♥️ 📓 (@joelgrus) June 26, 2019
This is the model more coffee shops need. I don't care about your whimiscal soft indie rock, I want a place to work that isn't my house. https://t.co/JbH3obuFYb
— Chris Albon (@chrisalbon) June 26, 2019
A few recent papers @NeurIPSConf also demonstrated that simple linear models can already achieve very competitive results on benchmark Mujoco RL tasks commonly used in the DeepRL literature:
— hardmaru (@hardmaru) June 26, 2019
NIPS 2017: https://t.co/MFezXDoK6c
NeurIPS 2018 (@beenwrekt): https://t.co/EzX9O6oLej
Motivation For Deep Learning https://t.co/eZ2bbpDzwV pic.twitter.com/rdB2dqZhj7
— Chris Albon (@chrisalbon) June 25, 2019
👋 Some super handy tips…
— Mara Averick (@dataandme) June 25, 2019
"Create Easy-to-Follow Technical Tutorials" 👩💻 @LittleMissData https://t.co/Gg9qBjaqk2 #rstats
Lots of academic SoTA results being shown from Mixup / CutMix / RICAP. But did you know it was originally (AFAIK) invented, in CutMix form, in a Kaggle competition 4 years ago, and shared in the discussion forum?
— Jeremy Howard (@jeremyphoward) June 25, 2019
Read Kaggle winning solutions! :)https://t.co/G3KVgg6D69
My favorite episode of Partially Derivative:
— Chris Albon (@chrisalbon) June 24, 2019
Should You Get A PhD? - https://t.co/SFyvamTRUG
Debugging is a superpower!
— Jenny Bryan (@JennyBryan) June 24, 2019
1. Requires & builds your confidence and tolerance of uncertainty.
2. Fully accept that something you believe to be true is definitely NOT TRUE.
3. Picture Ed Harris, in that vest, in Apollo 13: “let’s work the problem, people”. https://t.co/rhuBazRcbM
Me too! I've noticed a lot of passive-aggressive sub-tweeting and "jokes" on ML twitter, rather than directly and clearly stating a point and counter-point https://t.co/ClDzUUNsw5
— Jeremy Howard (@jeremyphoward) June 21, 2019
outsized influence on issues that matter to me. E.g. many in the govAI field are worried about memes like the "US-China Tech Cold War" bc they position tech development as one tied to civilizational supremacy, incentivizing powerful states to cut corners on things like AI safety.
— Jeffrey Ding (@jjding99) June 21, 2019
Look at these two celebrating how they coined the US-China Cold War meme. One I argue in this week's #ChinAI is not only flawed but extremely dangerous, fueling Cold War warriors who want to elevate existing tensions with China to civilizational conflict https://t.co/MdnV2gK5xC https://t.co/Nhtj33pGia
— Jeffrey Ding (@jjding99) June 21, 2019