Your state-of-the-art method is only state-of-the-art if other people can actually use it.
— hardmaru (@hardmaru) July 22, 2020
Your state-of-the-art method is only state-of-the-art if other people can actually use it.
— hardmaru (@hardmaru) July 22, 2020
.@kaggle Birdcall competition update:
— Radek Osmulski (@radekosmulski) July 20, 2020
You haven't heard from me in quite a while, because I didn't get any results worth sharing. I tried a couple of things to address domain shift (recalculating bn stats, https://t.co/4tb32YX6aw) but without access to test set at train time, pic.twitter.com/NneyJvo8qI
The evidence has been there for months
— Eric Topol (@EricTopol) July 20, 2020
Yet the lack of accepting by all to change behavior persistshttps://t.co/xHnFYQOM97 by @mccabe_caitlin pic.twitter.com/zH59BKP76h
An algorithm to determine medicaid had a bug in the code, and incorrectly cut needed care for patients with cerebral palsy. There was no way to contest or appeal these decisions. 10/https://t.co/qlU3bLjDcj
— Rachel Thomas (@math_rachel) July 20, 2020
Over time, I have come around and see the wisdom of his view. PCs are good at rejected the bottom 30-50% of papers that everyone agrees is not ready. Why get so hung up on the remaining 50%? We should just accept them all, and free up a ton of time for reviewers.
— Vijay Chidambaram (@vj_chidambaram) July 19, 2020
The JupyterLab real-time-collaboration effort funded from @cziscience just put up a website to coordinate information and activity across the community during the project! https://t.co/gnEPJ7AJM4
— Chris Holdgraf (@choldgraf) July 18, 2020
initial attempts: very impressive QA results (check out the coref in the gates questions!) but also has some glitches. pic.twitter.com/35hNLJWqWy
— (((ل()(ل() 'yoav)))) (@yoavgo) July 18, 2020
This is quite impressive. If you aren't astonished, you are a frog in a slowly warming pot. https://t.co/KsHqF4nWRm
— timoreilly (@timoreilly) July 18, 2020
Explainability: Why did the model deny the loan?
— Rachel Thomas (@math_rachel) July 17, 2020
Recourse: What can a person do to obtain a loan? Is there a feasible action they can take?
Recourse is more actionable for those impacted.
Talk here: https://t.co/1sfvv7pwkx @berkustun #PAML2020 pic.twitter.com/lNONaCn6uq
High predictive accuracy != good decision-making
— Rachel Thomas (@math_rachel) July 17, 2020
Few real-world problems collapse to maximizing accuracy for an already-coded variable. A computational system may be "optimized" in some sense, but usually not for what people really care about.@achould #PAML2020 #ICML2020 pic.twitter.com/uyCZJhGHVU
GPT-3 often performs like a clever student who hasn't done their reading trying to bullshit their way through an exam. Some well-known facts, some half-truths, and some straight lies, strung together in what first looks like a smooth narrative.
— Julian Togelius (@togelius) July 17, 2020
Excited to give a talk at ICML workshop on deploying machine learning systems pipeline in an hour.
— Chip Huyen (@chipro) July 17, 2020
Here are the slides for those who want to follow along. Feedback & questions much appreciated! https://t.co/9rM9BDD4Bn