Great walkthru from @CWerner76 showing how to create a 99% accurate guitar classifier, leveraging the latest fastai features and modern training trickshttps://t.co/QfarzvxnOa
— Jeremy Howard (@jeremyphoward) November 18, 2018
Great walkthru from @CWerner76 showing how to create a 99% accurate guitar classifier, leveraging the latest fastai features and modern training trickshttps://t.co/QfarzvxnOa
— Jeremy Howard (@jeremyphoward) November 18, 2018
I have to admit, I nearly injured my butt laughing at this @Quora thread. Thanks to Hosea Siu and Mo Nastri for their generous defense of our critical work. :) https://t.co/U9mGF0taCl
— Zachary Lipton (@zacharylipton) November 18, 2018
Great points made by @wpwalters on AI in drug discovery. At least for the ML part (building predictive models), there has been great progress and utility. @deepchem is a easy place to start for those new to the field. https://t.co/NRqVCb8php
— Vijay Pande (@vijaypande) November 17, 2018
This is beautiful! I love the strides @spacy_io team is making. It is obvious from all changes/features their deep understanding of the problems they are solving -- not just academically but also from a product usage POV. https://t.co/B2L71aW8uB
— Delip Rao (@deliprao) November 16, 2018
I have developed a sharp "too good to be true" skepticism about predictive models.
— Data Science Renee (@BecomingDataSci) November 16, 2018
Term for newbies to learn: Data Leakagehttps://t.co/DoebxVX4BH https://t.co/4ZXK9aCBRV
The man behind most hand-painted ski maps around the world, @JamesNiehues, is cataloging his life's work in a book https://t.co/CQyUQNrbNh
— Nathan Yau (@flowingdata) November 16, 2018
Data-driven reporting FTW https://t.co/DssqtYxzjO
— Julia Angwin (@JuliaAngwin) November 16, 2018
so... if this rate keeps up then around 2020 we'd be training ImageNet to 75% accuracy in 0.5 seconds :)
— Andrej Karpathy (@karpathy) November 16, 2018
Once upon a time, people thought the internet would help bridge the coastal-heartland divide, w/tech workers scattering across the U.S. Amazon's HQ2 spectacle shows how definitively that's not been the case: https://t.co/oSjZ02QOvR
— Te-Ping Chen (@tepingchen) November 15, 2018
📊📈📉👉Data Fallacies to Avoid — An Illustrated Collection of Mistakes that People Often Make When Analyzing Data: https://t.co/CxYW6FQT26 #abdsc #BigData #Analytics #DataScience #Statistics #DataLiteracy #DataEthics #DataScientists #StatisticalLiteracy pic.twitter.com/FZf1iQMZ1r
— Kirk Borne (@KirkDBorne) November 14, 2018
Everybody should watch this – literally!
— Emilio Ferrara (@jabawack) November 14, 2018
This is the best (by a long shot) coverage I have every seen on Russian influence campaigns! Operation Infektion: A three-part video series on Russian disinformation https://t.co/xP2WhSNZKy amazing work, @nytimes pic.twitter.com/7eSKGag2mx
😬 How to get those verbatim, backticked code chunks just right, so @xieyihui doesn't have to do it for you…
— Mara Averick (@dataandme) November 13, 2018
"The Two Surprisingly Hard Things about the Otherwise Simple Markdown"https://t.co/furHvAxwno #rstats