Benj Shapiro

Passionate about driving creative innovation in #MachineLearning, #genomics, #healthcare and #socialgood. PhD from Johns Hopkins U., AI/ML R&D @ GSK

[Clement Chow @ClementYChow]

Genetics!

[Benj Shapiro @benjdshapiro]

@tapairportugal I am following up, thanks

[Benj Shapiro @benjdshapiro]

Check it out to learn about our novel machine learning approach to integrating complex relationships between different data modalities + an interesting application to integrating epigenetics into an analysis of breast cancer patient data.

[Benj Shapiro @benjdshapiro]

Excited to share our paper on Bayesian Multi-View Clustering as part of F1000's collection on Machine Learning in Genomics! Huge thanks to @alexisjbattle for her support on this project. https://t.co/QMndpIRb4k

[Christoph Molnar 🦋 christophmolnar.bsky.social @ChristophMolnar]

I found it unexpectedly difficult to get into causal inference.

(still a beginner, I guess)

Here are a few insights that helped me in understanding causal inference. 🧵

[[email protected] @bb_apes]

you might be having a bad day, but just remember, there’s someone out there who has the cilantro soap gene and at least you’re not them

[Nicholas Coles, PhD @coles_nicholas_]

A year ago, a non-academic friend listened to a talk I gave. I thought it went great. My friend disagreed.

She said that academics are experts at making interesting stuff boring—and that we should all take a speech class.

So I did. And here are 6 most useful things I learned.

[Lior Pachter @lpachter]

SV: Using AI we have a cure or treatment for all diseases by 2100.

Healthcare: Great! But can you please fax us your healthcare records because we lost the index cards they were printed on.

[Barry @TheGrandBlooms]

All of mathematics is taught like someone explaining the rules of a board game that you're not playing yet

[Michael Retchin @MichaelRetchin]

Interested in biotech? Check out these free resources and events, people to follow, podcasts, and more.

All my Learning Biotech threads in one 👇

[Daniela Witten @daniela_witten]

barkus variance trade-off

[Benj Shapiro @benjdshapiro]

@TenanATC Agreed, I wish I saw concrete explanations of model failings + required next steps more often!

[Sam Altman @sama]

even people who really understand compounding consistently underestimate the power of compounding.

if you can get close to internalizing and acting on the true power of compounding and momentum, you will outperform almost everyone.

[Suchi Saria @suchisaria]

Hahahahahahahahahahahaha

[Cookie Monster @MeCookieMonster]

Da best thing you can cook is cookies. Dat's why they spell it dat way.

[Michael Baym @baym]

Science communication isn’t hard right? You just science and then open your mouth. Nothing else to it

[Korben Dallas Spence @austinrspence]

I have a job at a wine store between my PhD and postdoc. I’d love to make it more normal to not do science all the fucking time. I actually love it - I make more money than grad school, I leave work at work, and I’m learning more about wine and spirits. Honestly great.

[Marshall Burke @MarshallBBurke]

Are you an ML/AI whiz interested in working on problems that matter? Check out SustainBench, our new set of curated datasets + benchmarks for sustainability-related ML tasks. Great opportunity to improve measurement of poverty, ag, education, water, more https://t.co/3yPm8r7Yq1