Sooyoung Oh

Walking in a random world

[Sooyoung Oh @chitchattoe]

https://t.co/ChWFTqZGtI

[Lauren Oakden-Rayner 🏳️‍⚧️ @DrLaurenOR]

I was asked to comment on the recently published Google/Verily work on predicting cardiac risk from retinal images. Nice work, nice article. Most impressed that @jjvincent actually has a decent crack at defining AUC in a news article.

https://t.co/WWMIkFQbAD

[Reza Zadeh @Reza_Zadeh]

Missing data hinder replication of artificial intelligence papers. In 400 artificial intelligence papers presented at major conferences, only 6% included code, 30% included test data, & 54% included pseudocode.
https://t.co/iI0VwvGJnO

[Tianqi Chen @tqchenml]

With the emergence of new hardware, what are the challenges for deep learning systems? We try to answer it in our TVM paper https://t.co/0QKQ3Jg1Kr to provide reusable optimization stack for DLSys on CPU/GPU, and accelerators. I will give an oral presentation about it at #SysML

[New York Times Opinion @nytopinion]

I grew up in the Boston suburbs and inherited a stubborn New England refusal to acknowledge frigid temperatures. https://t.co/ztJAnrJoEO

[KING 5 News @KING5Seattle]

GeekWire: Time-lapse shows Seattle’s massive growth https://t.co/8Pm6qJHbqE

[Eric Horvitz @erichorvitz]

Recommend this clear, short piece in @sciencemagazine "What can machine learning do? Workforce implications" https://t.co/6hpVLZNAKG @erikbryn @compcomcon @PartnershipAI

[Mark Papadakis @markpapadakis]

https://t.co/yFH6rJWWsq : Incredible visualisations based on simple math formulas. Via @gmosx

[@fastml_extra @fastml_extra]

Vowpal Wabbit now has a companion: Neual Decision Twee https://t.co/kr7nveWvHE

[Nitish @StrongDuality]

New paper with @RichardSocher: "Improving Generalization Performance by Switching from Adam to SGD". https://t.co/AVSOER4aXw

[Lauren Oakden-Rayner 🏳️‍⚧️ @DrLaurenOR]

Great blog post from @DrHughHarvey on the barriers to using x-ray reports to build medical AI systems. If my last blog post was the "what", then this post addresses the "why" very well. https://t.co/3cvj12ayW0

[Guillaume Lample @ NeurIPS 2024 @GuillaumeLample]

We just open-sourced MUSE, our library to align embedding spaces in a supervised or unsupervised way, along with multilingual embeddings for 30 languages aligned in the same vector space, and 110 large-scale ground truth bilingual dictionaries: https://t.co/tCipxPJY2M

[hardmaru @hardmaru]

A nice summary of deep learning developments in 2017 by Eduard Tyantov. What's in store for 2018? https://t.co/hiIxh0HVPr

[Alex J. Champandard 🍃 @alexjc]

Wow! I can no longer distinguish between a computer generated voice and recording of a person. #TTS #generative #DeepLearning
Try the samples then the Turing test: https://t.co/8LNcaCGfLR

[Soumith Chintala @soumithchintala]

a birds-eye view into Facebook's datacenter infra
https://t.co/i8MaG0BTji

[Reza Zadeh @Reza_Zadeh]

Assessing image quality using Neural Networks, and further using the net to improve image quality. From Google. https://t.co/SOpwd0hfDm

[UtterlyInteresting @UtterlyInterest]

The letter A.

[Lauren Oakden-Rayner 🏳️‍⚧️ @DrLaurenOR]

I've spent several weeks exploring the ChestXray14 dataset, and I have some serious concerns with it.
https://t.co/52PSslbAh8

[Jeremy Howard @jeremyphoward]

This is exceptional work. Really exceptional. (I need to reread to fully understand the last section stoll.)

This piece shows why we need to teach deep learning to domain experts. Only a radiologist could have written this. https://t.co/z9YmXpa3Xz

[The New York Times @nytimes]

"I know I am not the only parent wondering if I can robot-proof my children’s careers" https://t.co/MuQ8osHExk