ML Research @ SmarterDx former @ColumbiaDBMI PhD, UC Berkeley EECS

Joined December 2017
Photos and videos
Check out our work published in EHJ today! This was a significant part of my dissertation and it was a joy to work with @PierreEliasMD and the rest of the team!
Today in #EHJ, we present the 1st time an AI model has outperformed radiologists in detecting cardiac pathology. ❤️‍🩹 Potential screen for #CHF w/ CXRs 🌁 Trained @ColumbiaMed validated @StanfordAIMI 🩻 Entire dataset (75k CXRs) released @ESC_Journals shorturl.at/cOT16
1
1
13
1,980
Shreyas Bhave PhD retweeted
Can you detect Stage B heart failure from a chest x-ray? Our ML model learns to detect LVH/dilated LV from CXRs done within 1 year of echo. Great work with @sabhave/@aperotte/@NirUrielMD. Currently validating at Stanford thx to @David_Ouyang & @mattlungrenMD! #ACC21
2
11
36
Shreyas Bhave PhD retweeted
India is in the throes of a horrendous COVID surge Horrendous They are struggling to get more people vaccinated We are sitting on 35-40 million doses of Astra Zeneca vaccine Americans will never use Can we please give or lend them to India? Like may be now? It'll help. A lot
1,150
14,385
75,633
Shreyas Bhave PhD retweeted
And further evidence that dispersed EMR - with no interoperability -- is a liability in a pandemic.. look at what the UK NHS database achieved in a few weeks. We should be ashamed ... x.com/bengoldacre/status/125…

OUR NEW PAPER. Factors associated with COVID-19-related hospital death in the linked electronic health records of 17 MILLION adult NHS patients. Largest study of its kind ever, anywhere. The power of UK / NHS data, realised. opensafely.org/outputs/2020/… medrxiv.org/cgi/content/shor…
14
66
192
Shreyas Bhave PhD retweeted
Make sure you add S104 to your Agenda in the #AMIA2019 app 📱!! I will be speaking tomorrow, Wed 11/20 ~9.30a, about how @ColumbiaDBMI used the #EHR to take a closer look at the urban #OpioidEpidemic. Preview here ➡️ youtube.com/watch?v=yeWflNo-…
9
10
Shreyas Bhave PhD retweeted
"Simple, Distributed, and Accelerated Probabilistic Programming". The #NIPS2018 paper for Edward2. Scaling probabilistic programs to 512 TPUv2 cores and 100 million parameter models. arxiv.org/abs/1811.02091 github.com/google-research/g…
8
206
716
Shreyas Bhave PhD retweeted
Are you interested in #causality or just want to leave #AMIA2018 with a bang? ‼️ Come to S115 in Plaza B on Wed, 11/6 from 1030-1200. I'll be presenting #NORA -- a novel #machinelearning model for #causalinference
9
6
Shreyas Bhave PhD retweeted
Vodka has been poured. Liberally.
1,034
1,356
18,915
Shreyas Bhave PhD retweeted
The disgrace of what has taken place with Brett Kavanaugh’s nomination will not be lost on the American people. We are mobilized from coast to coast.
1,555
5,054
28,779
Shreyas Bhave PhD retweeted
1/ A student just asked me why, if the p-value for a study is 0.04, we can't say the study has a 4% chance of being a false positive. First off, we definitely can't, even under idealized conditions - here's a brief thread as to why.
34
732
1,898
Shreyas Bhave PhD retweeted
This vote will allow senators to show once and for all where everyone stands on #NetNeutrality. We hope that we can make this bipartisan & our GOP colleagues will join us in standing up on behalf of average consumers & middle class families. x.com/netflix/status/9938873…

8 May 2018
Keep fighting for a free & open internet #netneutrality netneutrality.internetassoci…
102
480
1,277
Shreyas Bhave PhD retweeted
17 Apr 2018
Same charts, but side by side in a static image for easier comparison. Drug overdoses is the category between suicide and homicide.
3
55
136
Shreyas Bhave PhD retweeted
17 Apr 2018
I made this animation to show the avg share of cause of death from @CDCgov vs. avg share of @Google search trends and headlines from @guardian & @nytimes. The data was collected by Owen Shen et al. in a great post: bit.ly/ChartDeath Source code: bit.ly/DataVizCode
37
710
1,228
Shreyas Bhave PhD retweeted
Billions of lab tests are performed in the U.S. each year, and most are not stratified by demographic variables (e.g. race, sex, age). Should they be? John Ioannidis, @chiragjp, and I explore some of the thorny challenges in our new @JAMA_current piece: jamanetwork.com/journals/jam…
1
64
118
Shreyas Bhave PhD retweeted
31 Mar 2018
Whoa… 28 million people apply for 90,000 jobs with India's railway company buff.ly/2GFiUtl via @reddit

1
2
2
Shreyas Bhave PhD retweeted
The Trump Administration’s new census question about citizenship could lower the response rate, affect accuracy & is a clear attempt to undermine representation in states like CA where the immigrant population is larger than in any other state. mercurynews.com/2018/03/27/2…
2
2
6
Shreyas Bhave PhD retweeted
Is the (base)ball juiced? 53eig.ht/2F5D93H

28
112
256