- Twitter
RT @jasonhickel: Wow. Barcelona has declared a climate emergency that gets straight to the point: "The current economic model is bas…
- The voice of sadness is censored as sick. What if it’s sane? | Aeon Essays
- What A Day: Trial Like You Mean It | Crooked Media
- Bricks Alive! Scientists Create Living Concrete – The New York Times
- Twitter
RT @dhh: “DoorDash pays the average worker an astonishingly low $1.45/hour, after accounting for the costs of mileage and ad…
- Twitter
RT @SheldrickTrust: As one of the older female orphan #elephants at our Nursery, Tamiyoi (left) likes to offer a comforting & caring pr…
- Elizabeth Warren Is Waging a Full-Body Fight to Defeat Trump
- Twitter
RT @Susan_Hennessey: Here is the governor of Florida saying that “voting is a privilege.†In a democracy, voting is a right.
- The Pathos of “Cheer†and the Extraordinary Deceptions of Cheerleading | The New Yorker
- The Smartphone Has Ruined Space – The Atlantic
- Why Elizabeth Warren’s Gender Matters in the 2020 Election | Time
- Bring up the bodies: the retired couple who find drowning victims | Science | The Guardian
- Twitter
RT @MsPackyetti: Let’s be very clear: the effort to undermine the largest voting rights restoration in years in FL is an absolute mo…
- Twitter
RT @emilybazelon: For @NYTmag, I wrote about people who were exonerated, after convictions by non-unanimous juries, & the jurors who…
- Proposed Book Banning Bill in Missouri Could Imprison Librarians – PEN America
RT @SCBegley: From PEN America: A bill proposed in Missouri could send librarians to prison for letting kids access banned books
- Twitter
RT @SRobTweets: New blog post!
🤑 Build a fraud detection model with @TensorFlow
âš–ï¸ Understand why the model predicted fraud with… - An Adversarial Approach for the Robust Classification of Pneumonia from Chest Radiographs
While deep learning has shown promise in the domain of disease
classification from medical images, models based on state-of-the-art
convolutional neural network architectures often exhibit performance loss due to dataset shift. Models trained using data from one
hospital system achieve high predictive performance when tested
on data from the same hospital, but perform significantly worse
when they are tested in different hospital systems. Furthermore,
even within a given hospital system, deep learning models have
been shown to depend on hospital- and patient-level confounders
rather than meaningful pathology to make classifications. In order
for these models to be safely deployed, we would like to ensure that
they do not use confounding variables to make their classification,
and that they will work well even when tested on images from
hospitals that were not included in the training data. We attempt
to address this problem in the context of pneumonia classification
from chest radiographs. We propose an approach based on adversarial optimization, which allows us to learn more robust models that
do not depend on confounders. Specifically, we demonstrate improved out-of-hospital generalization performance of a pneumonia
classifier by training a model that is invariant to the view position
of chest radiographs (anterior-posterior vs. posterior-anterior). Our
approach leads to better predictive performance on external hospital data than both a standard baseline and previously proposed
methods to handle confounding, and also suggests a method for
identifying models that may rely on confounders. - Twitter
RT @brianschatz: I’m trying to be extraordinarily careful with my words here, so apologies if this is a bit stilted. 1)This @maddow…
- SecureMyEmail makes really private email surprisingly simple
- Two Jurors Voted to Acquit. He Was Convicted of Murder Anyway. – The New York Times
- Twitter
RT @k_mahnaz: Women in Data Science conference is coming back to Seattle. Call for proposal is open now. Also looking for sponsor…
- Twitter
RT @JamesGleick: Barr’s Justice Department has had all the Parnas evidence for months. Barr must explain why he suppressed it.
Digest powered by RSS Digest