Tobias Ross

Manager in the AI Garage @CapgeminiInDe

[Lena Maier-Hein (now mainly on LinkedIn) @lena_maierhein]

Agree that plain results tables are often not very meaningful?

🆕🆕🆕Then check out our "Beyond rankings" #MedIA paper: https://t.co/htsJIasLyE by @TobiRoss11 @annika_reinke1 et al.

We address the question which properties of an image characterize failure cases. 1/2

[Tobias Ross @TobiRoss11]

One cannot point this out often enough,
correlation != causality
https://t.co/uzVV5gb63Z

[Lena Maier-Hein (now mainly on LinkedIn) @lena_maierhein]

Are you excited about both, advanced ML methods and healthcare applications? Then this Postdoc position might be a great fit for you. In collaboration with Ulli Köthe‘s lab, we will work on uncertainty-aware image analysis with invertible neural networks. @MiccaiStudents

[Lena Maier-Hein (now mainly on LinkedIn) @lena_maierhein]

🪙🪙🪙This is the paper I was waiting for:
https://t.co/sYrkr71DZK

In full support of my hypothesis that most ML papers in healthcare are not very useful.

Thanks @random_walker and @sayashk as well as @michaelhoffman for the catch!

Highly relevant @bias_sig @MICCAI_Society

[Matthias Niessner @MattNiessner]

An important detail in pose optimization & bundle adjustment problems: each correspondence pair has 3 residuals (2 for re-projection errors)!

Merging these into a single residual collapses rows of the underlying Jaccobi matrix and 2nd-order / quasi Newton solvers won't work.

[Endoscopic Vision (EndoVis) Challenge @endo_vis]

Excited to announce #EndoVis #MICCAI2022 challenge with 6 diverse sub-challenges: https://t.co/tzSZrnWPK7
@SpeidelStefanie @lena_maierhein @DanStoyanov @annika_re @SophiaBano @CammaLab @WEISS_UCL @ICLHamlynRobots @IntuitiveSurg @vision_sharib @MICCAI_Society #biomedicalchallenges

[IMSY Lab (DKFZ) @DKFZ_IMSY_lab]

Congratulations on winning the GIANA polyp classifications challenge to Amine, Nuong and team! Keep up the great work🔥 @TobiRoss11 @SeidlitzSilvia @tj_adler. @helmholtz_ai @DataScienceDKFZ @ellislifehd

[Lena Maier-Hein (now mainly on LinkedIn) @lena_maierhein]

How to achieve patient benefit with #SurgicalDataScience?

Opinion of experts from 51 institutions based on multi-stage Delphi process. Just accepted by Medical Image Analysis https://t.co/NGQgvT802H with many contributors from @IPCAI_conf @ellislifehd and @MICCAI_society (1/3)

[Lena Maier-Hein (now mainly on LinkedIn) @lena_maierhein]

Congratulations (now) Dr. @TobiRoss11! Thanks for all your great personal and scientific contributions to @DKFZ_CAMI_lab!

[Tobias Ross @TobiRoss11]

@lena_maierhein @DKFZ_CAMI_lab It was (and still is) an honor being part of #DKFZ_CAMI_lab and thank you personally for all your support @lena_maierhein!

[Vlad Haltakov @haltakov]

Can you detect COVID-19 using Machine Learning? 🤔

You have an X-ray or CT scan and the task is to detect if the patient has COVID-19 or not. Sounds doable, right?

None of the 415 ML papers published on the subject in 2020 was usable. Not a single one!

Let's see why 👇

[MIT CSAIL @MIT_CSAIL]

9 Distance Measures in data science with algorithms.

via: @gp_pulipaka

[Mike Tamir, PhD @MikeTamir]

Biases in AI Systems - ACM Queue https://t.co/6XMbiLKI0a
#AI #MachineLearning #DeepLearning #DataScience

[Tobias Ross @TobiRoss11]

@derspiegel Dieser Artikel, der ohne ersichtliche Notwendigkeit über meine hochgeschätzte Kollegin @MinuTizabi so abschätzig und arrogant über ihr Leben sich herausnimmt (falsch) zu urteilen, ist absolutes Bildniveau. Pfui @derspiegel!

[Jane Lee @JaneLee17527424]

What is #Artificial #Intelligence #Surgery?
https://t.co/9YJTeOX7lu
@mahirozmen @ElyanEyad @IHPBA @AmCollSurgeons @salo75 @iss_sic @wis_italia @Abuhilal9Abu @konrad_karcz @ArnarGeirssonMD @CarlaPughMDPhD @Friebe_Michael @Belindadesimone @p_masca @SpeidelStefanie @teachmesurgery

[Maithra Raghu @maithra_raghu]

Do Vision Transformers See Like Convolutional Neural Networks?

New paper https://t.co/mxLCIRBRLy

The successes of Transformers in computer vision prompts a fundamental question: how are they solving these tasks? Do Transformers act like CNNs, or learn very different features?

[Ari Seff @ari_seff]

covariance vs. mutual information

[Chip Huyen @chipro]

Takeaway from @karpathy's CVPR talk:

The most successful ML projects in prod (Tesla, iPhone, Amazon drones, Zipline) are where you own the entire stack.

They iterate not just ML algorithms but also:
- how to collect/label data
- infrastructure
- hardware ML models run on

[Lena Maier-Hein (now mainly on LinkedIn) @lena_maierhein]

10,000 annotated images for the #SurgicalDataScience /@ipcai_conf community: https://t.co/c0pocXANxY. Thanks to @nature /@ScientificData for acknowledging the value of (open) curated data. A @DKFZ @NCT_HD @uniklinik_hd initiative with M. Wagner @TobiRoss11 et al. @MICCAI_Society

[Dave Keating @DaveKeating]

I keep seeing the line 'EU is having vaccine problems because it was too slow in negotiating contracts' repeated in 🇬🇧&🇺🇸 media.

I want to push back on this narrative because I think it's missing where real EU-level mistakes lie. Let's review what happened in past year (��1/17)