Mert R. Sabuncu 🤖🩻⚕️ @mertrory Twitter profile

Pinned Tweet

Mert R. Sabuncu 🤖🩻⚕️

1 month

There is one thing I keep hearing from NIH study sections, in paper reviews, or as questions at seminars: deep learning is not intrepretable and this makes DL problematic for medical imaging. This is flawed and here's something to read about this:

2

4

24

Last Seen Profiles

@PtfsFargo

@VmaniakJ

@wawafuu_

@rinyaa_san

@RobTopGames

@koorisgf

@zachosborne02

@LAPressClub

@aysha_2022

@CMaunoir13970

@singingsugar

@dakardingo

@bokeplokalmalam

@MicheleCHeyward

@Wizards_DnD

@robertlorraine2

@kibun_2

@yorrchive

@thiagonahueel

@Apple

@SarahDRasmussen

@IcelandMusic

@magnumpi

@bk4hyn

@KenJennings

@ourple_forever

@rauariza

@CoachTylerFunk

@RanaZah24710393

@STEREOLUUV

@NfinityLabs

@fooly444

@menzi_b2

@Ye9Tk6gebHfFCqu

@j_waggoner

@parkseohaam

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

Some advice I give my graduate students about how to keep on top of the literature in their field. A 🧵: 1/

136

2K

9K

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

Just received official word that I am promoted to full professor @Cornell effective January 1. I’d like to thank all my mentors, collaborators, and trainees for this amazing journey so far. Onward and upward.

78

14

666

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

It’s good to have a dedicated time (1 hour/day?) for finding papers, adding them to your read pile and “processing” them. 2/

7

11

636

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

For finding papers, my go-to tool is Google scholar. I try keyword searches, following citations, and author-based searches. I will also add random papers I encounter, say on Twitter, to my stack. 3/

7

9

550

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

That said, I strongly recommend having a system for processing the literature at a steady pace and not simply reactively. This habit can save a lot headache and improve the quality of your work. Fin.

32

14

497

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

Each paper added to a stack will be prioritized based on a quick skim of the title and abstract (level 0 understanding). I recommend a 3-stage prioritization. High, medium, low. High, for instance, will indicate a paper that is super relevant to my project. 4/

1

3

404

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

As an advisor to many grad students over the years, I’ve noticed that there are a few very important personal qualities that determine success. A brief 🪡 :

4

47

395

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

4 months

If you earn a PHD at an R1 US college, we should automatically grant you permanent residency.

15

23

341

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

5 years

We have gone too far with conferences. They are too big. Reviews are too noisy and not useful. We put too much weight on acceptance. They incentivize half baked flashy papers. They are bad for the environment. And they discriminate against those who cannot travel.

7

47

317

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

6 years

Randomness is so underrated in our culture. Yet it underpins evolution, intelligence, and every single facet of our lives. Some think it's your IQ or effort that determines success. The sobering truth is, pure luck is often the necessary condition that no one talks about.

8

87

291

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

You can aim to add and categorize ~10 papers a day. That should not take more than 10 minutes. Next, you move on to level 1 processing of your stack, starting from high priority papers. This level of processing should take about 10 minutes per paper and yield … 5/

1

4

285

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

3 years

In our recent paper in @ScienceAdvances , we use modern deep learning to show that the fMRI response to naturalistic stimuli is predictable across much of cortex w/ an accuracy close to the noise ceiling. Joint work with @meenakshik93 , @ngohgia @amykooz

2

62

246

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

7 years

A ConvNet paper (with back-prop and applied to a medical imaging problem nonetheless) from 1993(!). This is before LeCun and Bengio. And only 22 citations on scholar. Wow! RT @affrangi

(PDF) Computer-Assisted Diagnosis of Lung Nodule Detection using Artificial Convolution Neural...

PDF | Several output association methods and trainer imposed driving functions in conjunction with convolution neural network are proposed for general... | Find, read and cite all the research you...

www.researchgate.net

5

95

235

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

a basic understanding of the algorithm/model, problem, and data. The inputs and outputs, some modeling details, the literature context (what gap is being filled) will be part of this understanding. You can aim to process ~3 papers/day at this level. 6/

1

4

230

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

I recommend taking notes for level 1 processing: 3-4 sentences per paper and linking these to pdf and paper stack. Next up is level 2 processing, which can take 1-2 hours per paper. You can aim to do ~2-3 papers a week at this level, dedicating maybe 30 min/day to this task. 7/

2

217

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

For a given project (paper), you might need to process 100s of papers at level 1, dozens of papers at level 2, and 5-10 at level 3. 11/

1

3

204

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

Finally we have level 3 processing, where the aim can be replication or baselining. This entails an understanding that allows implementing the algorithm, deriving the theoretical results, and achieving the empirical results. You can probably do 1-2 of these a month… 9/

1

3

188

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

You can only decide if a paper should be level 2, after processing it at level 1 and based on how relevant/important it is for your project. Level 2 processing should lead to a good understanding of the theoretical and empirical details, and results. 8/

1

2

178

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

These recommendations and estimates are all subject and project-specific of course and should only be considered as rough guidelines. Moreover, it does not consider the broad reading that one should also pursue during their PhD and beyond. 12/

2

169

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

And the time you spend on this beyond the level 2 understanding can be considered as part of your research time (so not just literature review). 10/

1

2

154

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

5 years

Here's what a typical brain looks like as it ages. Warning, this is a little depressing to look at 😀 Consider this a teaser form a manuscript we plan to make public in a few weeks - @AdrianDalca

6

58

152

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

The 🚠 commute to ⁦ @cornell_tech ⁩ never gets boring.

5

145

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

5 years

There must be a fundamental reason relevant to learning for biological neurons to be so diverse.

2

24

126

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

9 months

Reference letter from a US Prof: "This is the best student I have ever had. If you don't hire them, you will regret it for the rest of your life."

9

1

129

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

3 years

I completed the review of a paper. It took me ~5 hours. And a lot of mental effort. In the middle of a pandemic, where every second and ounce of energy is valuable (parents of little kids, you know what I mean). And I did this for free. OA fees or paywalls don't pay for this.

1

10

123

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

5 years

We are gearing up to launching our virtual seminar series in Machine Learning in Medicine - aiming to minimize carbon footprint + speaker burden. Sending out speaker invitations shortly. Will announce line-up in real time. Watch this space and spread the word.

4

33

120

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

3 years

The bar for faculty interviews each year:

1

7

102

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

3 years

I am so proud to win ⁦ @CornellEng ⁩’s undergraduate teaching award - and this was my last opportunity before my recent move to ⁦ @cornell_tech ⁩. ⁦ @CornellECE ⁩

11

0

102

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

4 months

We should educate grad students and post docs about the finances of running an academic research group. Too many, I find, are absolutely clueless about where the funds come from and what they pay for.

5

94

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

Some advice for giving good academic talks. A short 🧵:1/

5

11

91

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

Sometimes the view from our classroom at @cornell_tech can be distracting.

8

3

81

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

We have an exciting faculty job opportunity at the intersection of AI and medical imaging at Cornell, in NYC. Please consider applying and help spread the word.

Research Faculty Position

The Department of Radiology at Weill Cornell Medicine invites applications for faculty positions at all levels. This search is part of a strategic initiative to build up a research ecosystem in Art…

cra.org

1

33

78

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

Happy first day of classes @cornell ⁦ @cornell_tech ⁩ ⁦ @CornellEng ⁩

0

1

77

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

Are you an incoming ⁦ @CornellEng ⁩ PhD student interested in machine learning and healthcare? We have great opportunities on the ⁦ @cornell_tech ⁩ campus in NYC, minutes away from ⁦ @WeillCornell ⁩. Reach out/spread the word!

1

19

77

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

1 year

A brief history of scientific publishing. A 🪡: 1/

5

21

75

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

6 years

Our review of Machine learning in resting-state fMRI analysis

Machine learning in resting-state fMRI analysis

Machine learning techniques have gained prominence for the analysis of resting-state functional Magnetic Resonance Imaging (rs-fMRI) data. Here, we present an overview of various unsupervised and...

arxiv.org

0

21

73

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

Are you a PhD or Masters student at @cornell_tech , with a background in machine learning? I’m offering an advanced topics class on interpretable and explainable ML this spring. Please help spread the word! It’s going to be fun 🤩

2

5

73

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

It’s good to remind oneself that scientists (even great ones) can be profoundly wrong and we are all prone to our own biases and flaws. Look at this gem from 65 years ago 👇

5

8

73

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

6 years

In other news: I was awarded an NSF CAREER:

17

0

66

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

5 years

Machine learning performance: Training data vs validation data vs real world data

Adi

@IDdocAdi

5 years

LIFE IS THE BEST MODULE Online modules Real life I I I I I I V V V #stewardmeme

14

137

500

0

11

62

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

11 months

Everyone agrees that interpretability or explainability is critical in ML for medical imaging (MLMI). In our brand new paper, we formalize and present the needs and elements of Interpretability in Machine Learning for Medical Imaging:

A Framework for Interpretability in Machine Learning for Medical Imaging

Interpretability for machine learning models in medical imaging (MLMI) is an important direction of research. However, there is a general sense of murkiness in what interpretability means. Why...

arxiv.org

0

12

61

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

5 years

Progression in ML (1): *Linear models w/ carefully selected variables can predict *Linear models w/ lots of variables predict better (but avoid overfitting) *Kernels can help turn linear into nonlinear *Nonlinear models do well too *Nonlinear models on raw data do great

1

22

60

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

5 years

Researchers do virtually all the heavy-lifting in scientific publishing. Largely, for free. And what do we get in the end? Content that readers pay to access or authors pay to publish. Plus, authors often lose the copyright to their own work. @MELBAJournal will have none of that.

2

13

61

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

3 months

Missing variables are ubiquitous in real-world data, especially in healthcare. We present Knockout: a theoretically grounded, easy-to-implement and very effective way to handle missingness in machine learning (ML). 1/n

1

10

57

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

5 years

The predictability ceiling is often ignored in mainstream ML research. Every prediction problem has an upper bound for prediction—the Bayes-optimal performance. If you don’t have a good sense of what it is for your problem, you are in the dark.

2

12

58

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

5 years

Our paper is finally published in NeuroImage. Ensemble learning with 3D convolutional neural networks for functional connectomes. … Pre-print:

Ensemble learning with 3D convolutional neural networks for...

The specificty and sensitivity of resting state functional MRI (rs-fMRI) measurements depend on pre-processing choices, such as the parcellation scheme used to define regions of interest (ROIs)....

arxiv.org

1

24

57

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

11 months

Hanging out at MICCAI with the next generation of awesome researchers from ⁦ @Cornell ⁩

0

58

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

Explainable ML

1

3

56

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

3 years

Another one for your Christmas reading list. ⁦ @ngohgia ⁩ ⁦ @amykooz ⁩ ⁦ @meenakshik93 ⁩ Keith Jamison and I show how one can predict individualized task activation maps based on resting state fMRI and using customized deep learning.

1

19

57

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

5 years

We are planning a "green" seminar series in "machine learning in medicine" at Cornell, where the speakers will give their talks virtually. Goal: minimize carbon footprint & burden on speaker. Hope to attract great speakers. Spread word please. And all suggestions are welcome.

3

9

56

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

4 years

We need *high-quality* virtual conferences that are <$50 to register for anyone. If anyone is interested in doing this for ML/Computer Vision/Medical Imaging ping me.

4

56

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

3 years

I hate it when journals make you do a ton of work (formatting, following strict word/page limits, filling out forms, suggesting reviewers, editors, etc) just to submit for review - and then desk reject you the next day.

4

3

55

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

1/ stick-to-iteveness: research is full of road blocks and can at times be unpleasant. Getting concrete results often demands persistence, creativity in problem solving, clarity in your vision and conviction.

1

55

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

A typical applied ML paper these days: we ran experiments on 10 different datasets, using 20 different architectures, each with 100B learnable parameters, using 1000 GPUs for 20 weeks straight and found that a mini batch size of 8 works fine most of the time.

2

1

54

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

I know academics tend to complain about teaching and grant writing. But, I can’t imagine better incentives for organizing one’s thoughts and learning new material.

5

2

54

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

3 years

“Linear models are more interpretable and (thus) preferable” is one of science’s most insidious beliefs.

5

6

54

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

4 years

From Connectomic to Task-evoked Fingerprints: Individualized Prediction of Task Contrasts from Resting-state Functional Connectivity. Our paper w/ @ngohgia @AmyKuceyeski @meenakshik93 :

From Connectomic to Task-evoked Fingerprints: Individualized...

Resting-state functional MRI (rsfMRI) yields functional connectomes that can serve as cognitive fingerprints of individuals. Connectomic fingerprints have proven useful in many machine learning...

arxiv.org

0

16

53

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

1 year

First three stages of a research project: 1) irrational exuberance (no one has thought of/observed this before), 2) denial (oh that paper is close but not the same), 3) acceptance (how can I pivot).

1

2

53

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

6 years

The different goals ppl have in mind when they say "to interpret a model". 1)What can I change to alter outcome?, 2)what part of this data-point influenced this decision (most)?, 3)which historical (training) data-points influenced this decision?, 4)show me typical failure modes

1

17

50

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

5 months

We should be discouraging the rat race mentality, not encouraging it. We should be incentivizing deep understanding, not superficial creativity. We should be emphasizing long-term planning, not short-term results.

3

6

52

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

11 months

Are you a PhD researcher in the field of computational medical imaging? Are you looking for faculty jobs? We, at @WCMRadiology , are hiring. Please read on. 🧵1/

1

23

49

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

Thanks everyone for the kind words. I wanted to say two more things. First, I can’t be happier to call @cornell my home. This is an amazing place and I’m very excited about building up Engineering in NYC. Second, I have many role models that inspire me and hope to thank some day.

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

Just received official word that I am promoted to full professor @Cornell effective January 1. I’d like to thank all my mentors, collaborators, and trainees for this amazing journey so far. Onward and upward.

78

14

666

4

2

50

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

5 years

Victor, an undergrad in our lab, just created this. 🤣

2

0

49

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

I watched the GWAS community obsess over statistical power and grow their sample sizes to the millions, using univariate tests to detect associations between single genetic markers and complex phenotypes. We now have hits that “explain” a tiny fraction of the variance.

2

9

49

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

5 years

We will make an exciting announcement - probably next week. An open access, streamlined, high quality publication venue.... for our modern times.

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

6 years

We are planning our UnJournal movement, which can help free us from the constraints of conventional scientific publishing models. I dream of the day when "UnJournal" catches on as a word and represents a new way of scientific dissemination :) Let's see....

2

4

21

2

4

49

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

7 months

@WeillCornell @PeterHotez Congratulations Dr @PeterHotez . Your commitment to research and scientific communication is commendable.

0

5

49

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

Good to see ⁦ @DanielRueckert ⁩ at ⁦ @cornell_tech ⁩ today!

0

1

48

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

4 years

My 9 yo son showed my how he has been circumventing screen time limits on his ipad: by starting a zoom meeting and screen sharing! OMG!

3

1

46

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

1 year

I met a few prospective PhD students today and got very excited about all the amazing future work that we are going to get done! I ❤️ this job!

0

47

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

2/ good communication: you can’t have an impact unless you convince other people to be inspired by your idea and build on it. Your work is as valuable as you can communicate it - in writing and verbally.

1

46

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

3 years

This is shameful. Scientific publishing companies that have historically enjoyed outlandish profit margins, largely on the back of the hard work of academic researchers who are often supported by tax payer money, should not be able to get away with this.

Nature Neuroscience

@NatureNeuro

3 years

In this scenario, the cost of publication is covered by an Article Processing Charge (APC) paid at the time of publication. The APC for Nature Neuroscience in 2022 is €9,500/US $11,390/£8,290.

382

207

449

3

0

47

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

11 months

Semantic interpolation in latent space:

1

4

46

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

6 years

Nips sold out in less than 40 min. I repeat, an academic conference is sold out at about the same rate as a Radiohead concert. If this ain’t an indication of an ML bubble, what is?

3

14

47

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

6 years

Look what I got for lab members to read 😀 hope this will inspire some good and thoughtful science. ⁦ @yudapearl ⁩ #Bookofwhy

1

45

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

5 years

Please stop using convolutional layers for problems with no shift-invariance/equi-variance. If (1) you do not understand what this means, (2) you're at Cornell and can take classes - consider taking ECE 4250 (DSP) in the spring.

4

0

46

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

1 year

Many top-tier scientific journals are published by companies that care about revenue. That’s why we have the obscene dichotomy: scientific articles behind unethical paywalls or outrageous ($1000s) article processing fees charged to authors. There is another way…

1

9

45

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

5 years

I’ll make it very simple: any high profile applied (eg medical) machine learning paper should evaluate the tech in a real world scenario, prospectively. If that’s not the case, it doesn’t deserve public and widespread attention. Period. Journalists, get that?

3

4

44

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

5/ maintaining a good habit of regularly reading: keeping on top of the literature is something you need to constantly work at. Fin. See:

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

Some advice I give my graduate students about how to keep on top of the literature in their field. A 🧵: 1/

136

2K

9K

1

0

44

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

I will be recruiting multiple PhD students from the ECE, CS, and BME programs in @CornellEng . Application deadlines are soon! If you’re interested in the intersection of machine learning, computer vision and biomedical imaging, consider mentioning that you’d like to work with me.

4

14

45

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

4 years

I'm looking for a post-doc w/ a strong machine learning and computer vision background who'd be interested in working on a prostate cancer imaging problem. This would be in New York City, on the @WeillCornell and @cornell_tech campuses. Please DM with leads. Email me w/ CVs.

0

27

43

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

5 years

My dad would have been 69 today. Happy birthday dad! Lots of ♥️ to all of you who have lost someone close.

1

0

42

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

Super happy to share our latest paper- amazing work largely done by ⁦ @batuhan7karaman ⁩ : Machine learning based multi-modal prediction of future decline toward Alzheimer’s disease: An empirical study | PLOS ONE

Machine learning based multi-modal prediction of future decline toward Alzheimer’s disease: An...

Alzheimer’s disease (AD) is a neurodegenerative condition that progresses over decades. Early detection of individuals at high risk of future progression toward AD is likely to be of critical...

journals.plos.org

2

9

43

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

5 years

I've decided to tweet this once every 3 years - until I feel the hype is over. 🙃

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

8 years

Replace "information theory" with "deep learning" or "machine learning" etc

0

10

15

2

10

41

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

5 years

If you know anyone who has a track record of working at the intersection of neuroimaging, machine learning, genetics, and Alzheimer’s, and they are looking for a postdoc, can you hook me up? Thx!!

1

34

42

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

3 years

Today will be my last lecture to @CornellEng undergraduates, as I am moving from Ithaca in a few weeks. You can find me hanging out on the @cornell_tech campus in NYC soon.

1

0

42

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

9 months

Reference letter from non-US Prof: "This person took my class and received a respectable B-. We wrote a semi-decent paper together on this topic. I think they are qualified for the position you are considering them for."

2

1

43

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

My attempt to describe CT to an eleven year old

4

2

42

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

3 years

When, in grad school, I first started reading scientific papers, my aim was to absorb/understand. It took me years to change my mindset and have a critical perspective. Now, I try to teach my students that one should question everything in a paper. Even best papers are flawed.

1

5

41

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

3 years

Every time I try logging in on the NIH website 🤣😅

Steve Stewart-Williams

@SteveStuWill

3 years

56

2K

14K

1

42

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

This is a paper we worked on for almost 3 years! Machine learning can aid in prediction of IDH mutation from H&E-stained histology slides in infiltrating gliomas | Scientific Reports

Machine learning can aid in prediction of IDH mutation from H&E-stained histology slides in...

Scientific Reports - Machine learning can aid in prediction of IDH mutation from H&E-stained histology slides in infiltrating gliomas

www.nature.com

1

10

42

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

6 years

Many training datasets have noisy labels, especially in biomedical applications. Zhilu Zhang and I propose an easy-to-implement loss function that can make your classification model more robust to noise. The paper is under review.

Generalized Cross Entropy Loss for Training Deep Neural Networks...

Deep neural networks (DNNs) have achieved tremendous success in a variety of applications across many disciplines. Yet, their superior performance comes with the expensive cost of requiring...

arxiv.org

0

17

41

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

4/ flexibility: as any good list, this one introduces a contradiction - in particular with (1). While persistence is important, sometimes you are at a dead end, and recognizing that and changing course is critical.

1

0

40

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

2 years

IMO, this way of thinking can deplete resources, kill scientific creativity, and reveal very limited insights with little practical value. I urge the neuroimaging community not to fall into this trap.

2

1

41

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

5 years

Open access is a must in scientific publishing. Another must is high quality, well curated content. Finally, there is no need for high publication fees, which amplify inequalities. @MELBAJournal will be built on these principles.

1

8

41

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

3 years

I said this before and I'll say it again: adversarial robustness/learning/GANs/etc are way over-hyped.

3

2

38

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

1 year

We present a very simple ML idea for Comparing Longitudinal Medical Images - will be presented at ⁦ @midl_conference ⁩

Learning to Compare Longitudinal Images

Longitudinal studies, where a series of images from the same set of individuals are acquired at different time-points, represent a popular technique for studying and characterizing temporal...

arxiv.org

1

8

40

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

7 years

Advice for new faculty from MIT Physics:

1

22

40

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

4 years

Intelligence plays dice. In this Perspective Article, I argue that stochasticity is essential for machine learning. Pre-print:

Intelligence plays dice: Stochasticity is essential for machine learning

Many fields view stochasticity as a way to gain computational efficiency, while often having to trade off accuracy. In this perspective article, we argue that stochasticity plays a fundamentally...

arxiv.org

2

9

40

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

4 years

For 3+ years now, since I moved to Cornell, I have been holding a ~weekly journal club. Many of the papers we read shape our views and research. If you are interested in taking a glimpse of what we have read or are reading, here you go:

2

37

Mert R. Sabuncu 🤖🩻⚕️

@mertrory

3 years

Most papers in machine learning these days...

Oded Rechavi

@OdedRechavi

3 years

Lab meeting when you don’t have new results

134

3K

18K

0

7

39