Deutsch
Walter de Back ๐จ๐ปโ๐ป
479 posts
Walter de Back ๐จ๐ปโ๐ป
@wdeback
computational pathology โข deep learning โข medical imaging โข @AIgnostics
Germany ๊ฐ์
์ผ Temmuz 2014
288 ํ๋ก์356 ํ๋ก์
Really important and resourceful paper about losses for medical image segmentation sciencedirect.com/science/articlโฆ
English
Our story with @morpheus_lab has been more fortunate, but I can fully relate to this. Research software engineers #RSE are essential parts of science and need to be treated as such. Kudos to @petebankhead for this story and to his fantastic software @QuPath!
Pete Bankhead@petebankhead
A small number of people know the real background story to @QuPath, but most don't. I didn't plan to ever tell it publicly, until a Google Alert today caught my eye. A thread about open science & academia ๐ (1/n)
English
Job alert: Postdoc position in Biomedical Deep Learning available @IMB_TUDresden - my former employer. Task: Develop innovative machine learning methods on real-world medical data. Check it out.
tu-dresden.de/med/mf/imb/dasโฆ
English
Walter de Back ๐จ๐ปโ๐ป ๋ฆฌํธ์ํจ
โ ๏ธ PRINCIPAL INVESTIGATORS WANTED โ ๏ธ
@HelmholtzMunich launches the Institute of AI for Health #AI4H, a European #AI #ML hub aiming to transform the future of medicine.
Hear more about this unique opportunity from director @fabian_theis & APPLY NOW!
๐ bit.ly/joinAI4H
English
Walter de Back ๐จ๐ปโ๐ป ๋ฆฌํธ์ํจ
Join us on the 8th of October for the virtual PANDA workshop at @MICCAI2020, with the first results of our challenge on prostate cancer grading: 11.000 biopsies, 1.000 teams & pathologist-level AI models! More info: panda.grand-challenge.org
English
Walter de Back ๐จ๐ปโ๐ป ๋ฆฌํธ์ํจ
@irinavlh, @DaniloJRezende and I are presenting a tutorial at ICML on 'Representation Learning Without Labels'.
Jul 13, 9 AM to 12 PM and 7 PM to 10 PM (BST)
icml.cc/virtual/2020/tโฆ
icml.cc/Conferences/20โฆ
Drop by to find out what Plato has to do with VAEs, GANs and SimCLR!
English
Walter de Back ๐จ๐ปโ๐ป ๋ฆฌํธ์ํจ
Working at the intersection of #AI and #Health? Here some overview articles I found useful:
(1) Comparison of deep learning performance against health-care professionals (Lancet Digital Oncology 2019):
thelancet.com/pdfs/journals/โฆ
(Note: major criticism of validation standards)
English
English
Personally, I don't think the pretrained *H&E* #StarDist model from @martweig and @uschmidt83 has any right to be able to detect nuclei here.
GIF
English
Walter de Back ๐จ๐ปโ๐ป ๋ฆฌํธ์ํจ
Interested in researching AI methodology from a geometric view point?
We have a vacancy for 2 PhD positions in the Amsterdam Machine Learning Lab (@wellingmax, @UvA_Amsterdam) on geometric machine learning for medical image analysis! (Open until May 6)
uva.nl/en/content/vacโฆ
English
Walter de Back ๐จ๐ปโ๐ป ๋ฆฌํธ์ํจ
Interested in doing a PhD at the intersection of machine learning, biological image analysis, and computational microscopy?
Then join my group @epfl in beautiful Lausanne, Switzerland - 1 week left to apply!
epfl.ch/education/phd/โฆ
Please RT
English
interesting use of Mรถbius transformations for data augmentation
Torbjรถrn Lundh@torbjornlundh
Data augmentation with M\" obius transformations - goo.gl/scholar/zDQixH #ScholarAlerts
Romรขnฤ
@JeanKossaifi @nvidia @AnimaAnandkumar @NvidiaAI congrats! will you be able to continue to develop or maintain your fantastic #tensorly library?
English
Really excited to be joining @nvidia as a senior research scientist, working in #ML and #tensor methods with @AnimaAnandkumar and the wonderful team there @NvidiaAI
English
@AlexGFletcher @adrs0061 @RaisingFloors @morpheus_lab @morpheus_lab is designed for cell-based modeling but also handles simpler ODE and PDE models. Not exactly drag and drop, but no programming needed.
Check out the Morpheus courses: morpheus.gitlab.io/courses/
Or the โmy first modelโ section of this tutorial: gitlab.com/wdeback/Workshโฆ
English
@adrs0061 @RaisingFloors @morpheus_lab Thanks Andreas, didnt realise it handled simpler models; will check it out.
English
@LittleMissData @JanHasenauer as a matplotlib user, youโre making me jealous ...
English
agree, nvtop is very useful. github.com/Syllo/nvtop
Peter Steinbach@psteinb_
Just discovered a #gemforgeeks: a htop like monitor of @NVIDIADC #GPUs. Super helpful to see that @TensorFlow is allocating RAM from all devices it can grab, but apparently only uses 3/4 GPUs. Thanks #Syllo: github.com/Syllo/nvtop #OpenSourceSavesTheWorld
English
On the current state of (digital) pathology in Germany: Who still wants to count cells?
sueddeutsche.de/gesundheit/medโฆ
English
Walter de Back ๐จ๐ปโ๐ป ๋ฆฌํธ์ํจ
Growing Neural Cellular Automata -- A new Distill article by @zzznah, @RandazzoEttore, @eyvindn and @drmichaellevin
distill.pub/2020/growing-cโฆ
English
Hereโs a blog reporting on our hackathon work predicting the Elbe river flow and uncertainties using ConvLSTMs. Written by Lennart Schmidt.
Peter Steinbach@psteinb_
@sunitachandra29 Thanks, I much appreciate the praise. For those looking for an in-depth report of one of the "most accurate" teams, see the prequel blog post on high tide predictions of the Elbe river with #AI mlcdresden.gitlab.io/pages/2019/10/โฆ @wdeback @kiravino
English
@NatureCellBio Congratulations to all collaborators, and in particular to Mehmet Saรงma (@stemcellniche), Johannes Pospiech and Carolina Florian!
English
@NatureCellBio Here's the sharable link:
rdcu.be/bV5rV
English
Out now: our paper on ageing of the hematopoetic stem cell niche @NatureCellBio.
nature.com/articles/s4155โฆ
English