Profile picture
, 3 tweets, 5 min read
My Authors
Read all threads
Tree-based machine learning to explain and monitor #AI model performance in the hospital setting
nature.com/articles/s4225… @NatMachIntell by @suinleelab @uwengineering and team
Also preprint for access: arxiv.org/abs/1905.04610
@NatMachIntell @suinleelab @uwengineering and explainable trees to better discern
—mortality risk factors in US
—identify subgroups with shared risk
—define interactions among chronic kidney disease risk factors
@NatMachIntell @suinleelab @uwengineering And a nice explainer of the explainer!
nature.com/articles/s4225…
by @wojciechsamek @FraunhoferHHI
a succinct commentary on the above paper
Missing some Tweet in this thread? You can try to force a refresh.

Enjoying this thread?

Keep Current with Eric Topol

Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

Twitter may remove this content at anytime, convert it as a PDF, save and print for later use!

Try unrolling a thread yourself!

how to unroll video

1) Follow Thread Reader App on Twitter so you can easily mention us!

2) Go to a Twitter thread (series of Tweets by the same owner) and mention us with a keyword "unroll" @threadreaderapp unroll

You can practice here first or read more on our help page!

Related hashtags

#AI

More from @EricTopol see all

Profile picture
Eric Topol
@EricTopol
There's "precision medicine" and "rabbit hole medicine"
This report represents the latter. How to collect an avg $8,000/person, do many tests to get lots of abnormal findings (for more testing) and "health outcome and benefits were not measured" pnas.org/content/early/… @PNASNews
@PNASNews My colleague @evanmuse and I wrote about this @TheLancet recently thelancet.com/pdfs/journals/…
"High-dimensional data for a human being is an alluring
yet unproven concept; "there is no validation..in the clinic"
@ScrippsRTI @scrippsresearch
@PNASNews @EvanMuse @TheLancet @ScrippsRTI @scrippsresearch Here's the previous report from the same group on "precision medicine" which I label (IMO more precisely) "rabbit-hole medicine." Of 209 people, >33% underwent additional medical imaging due to incidental findings pnas.org/content/pnas/1…
@PNASNews
Read 4 tweets
Profile picture
Eric Topol
@EricTopol
My 1st #AI-guided echocardiography today (fun & by democratizing acquisition it will change the field)
What does that mean?
You put the transducer on the patient's chest ->the AI tells you how to properly position it and then automatically captures video images when spot on.
Another view I obtained (short-axis) letting the #AI guide me w/ the probe position and auto-loop capture the image.
Algorithms by @CaptionHealth (I have no COI or relationship)
Software in development for interpretation and for the rest of the body's organs (except the brain😉)
@CaptionHealth + some solid #AI work on echo interpretation just published a few days ago @Nature_NPJ Digital Medicine #openaccess
nature.com/articles/s4174…
by @GhorbaniAmirata @abidlabs @bryandhe
@jonc101x @HeartBobH @euanashley @David_Ouyang
@james_y_zou @StanfordMed @StanfordDeptMed
Read 3 tweets
Profile picture
Eric Topol
@EricTopol
Over the past decade, the brilliant and imaginative Randall Munroe @xkcd @xkcdComic has introduced 7 smartphone versions, each with advancing sophistication :-)
(After 6.0, he's been quiet on this for awhile) 1/3
@xkcd @xkcdComic Here's my (updated) projection for the medicalized smartphone (red ✓ = achieved) 2/3
@xkcd @xkcdComic I haven't yet seen any @xkcdcomic smartwatch versions, but here's my conception of where we are headed
(red ✓= already demonstrated or in prototype) 3/3
Read 3 tweets

Related threads

Profile picture
Bradley Love
@ProfData
New paper w @BDRoads, "Learning as the Unsupervised Alignment of Conceptual Systems". Supervised learning tasks can be solved by purely unsupervised means by exploiting correspondences across systems (e.g., text, images, etc.). 1/5 nature.com/articles/s4225…
How can supervised tasks be solved without supervision? Concepts have unique signatures that hold across unsupervised systems. E.g., The words pen and pencil pattern similarly in text corpora and are also visually similar. Likewise, both are different to lions, cars, etc. 2/5
Unsupervised alignment only gets better with more concepts and systems. Each concept's signature becomes more unique the more concepts there are in each system (see below). Each system provides a slightly different viewpoint on reality such that adding more systems helps too. 3/5
Read 6 tweets
Profile picture
Misha Denil
@notmisha
Reproducibility is the wrong objective for Machine Learning. Reproducibility is key for science, but ML is not a science. nature.com/articles/d4158…
ML is a community of makers. Instead of reproducibility we should optimize for mindshare.
You want people to build on your work, so optimize for that. Sometimes releasing code is the way to make that happen (e.g. DCGAN), but sometimes influence comes in other ways (e.g. AlphaGo).
Read 4 tweets
Profile picture
Medical Science and Technology
@MedicalScitech
UN FUTURO MAGICO. PECCATO PER GLI UMANI (SCORDATE L’EUROZONA 2).
(2017-03-27)
» paolobarnard.info/intervento_mos…

"Qui l’arrivo di VR e di AR compiono come mai immaginato prima, né sognato prima, il sogno di Edward Bernays"

VR AND AR PUSHING CONNECTIVITY LIMITS
» qualcomm.com/media/document…
Researchers at the Samsung #AI Center have produced a system that can recreate realistic motion from just a single image of a person
» arxiv.org/pdf/1905.08233…

"We believe that telepresence technologies in AR, VR and other media are to transform the world"
Graphene could soon make your computer 1000 times faster
» weforum.org/agenda/2017/08…

“If you want to continue to push technology forward, we need faster computers to be able to run bigger and better simulations for climate science, for space exploration, for Wall Street”
Read 9 tweets
Profile picture
Alexis Verger
@Alexis_Verger
#365papers Let's try this for 2019, highlighting 365 papers (does not necessarily mean 1 paper per day) - A thread
#365papers 1⃣ The thread started with a bang - A must read opinion by @MarcHalfon on the fundamental question of how to define an enhancer cell.com/trends/genetic…
#365papers 2⃣ citrullination by PADI2 of RNA Pol II CTD R1810 regulates the transcription of highly expressed genes by mediating the interaction with P-TEFb cell.com/molecular-cell…
Read 367 tweets

Trending hashtags

#Rust #equivalentieprincipe #SDoH #China #tfjs #France #DeepLearning #DataValidation #StrongWomen #dataset #LLPS #ai #Songar #doodle #AI #womeninscience #IA #CRISPR #QuantumAI #data4good #ML #StyleGAN2 #LibraInDC #halloween #AIEthics
Follow Us on Twitter!

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just three indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3.00/month or $30.00/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!