Too bad I threw away a bag of RAM sticks, lost a fortune...

Another grant approved, makes 4 in 2025

When the paper you spent 6 years on finally gets a citation

Im now looking for a postdoc in AI-Driven Precision Orthopaedics. We have huge datasets and plenty of GPUs. liu.se/en/research/...

Just reviewed a paper which shared data and code on Google drive 🤔

Fina minnen

How would you react as a reviewer if you review a paper with 4 types of data from two groups, and the group comparison using each type of data is done in a (slightly) different statistical way?

How common is it to do the data analysis as a PI?

Did not make it to step 2 for ERC, meaning that part B2 (14 pages) was never read 🥲

Grant funded, more than 4 years since the last big grant. More than enough to hire a PhD student.

Trying to get funding for AI research in medicine Technical funding agencies: you have not published in NeurIPS and ICML Medical funding agencies: you have not published in medical journals

How do you deal with grant rejections when the researchers who got grants have substantially weaker CV than yourself?

Lets gooo

Interesting conference at Stanford

Why can I still not do polls on bluesky?

Another 3 grant rejections, without a single word of feedback

Are you working with diffusion MRI? Have a look at this analysis-ready and quality controlled dataset of 2,747 participants aged 5–21 years. I was one of the researchers who performed quality control for at least 3000 images. www.nature.com/articles/s41...

Attended this great workshop on reproducibility and asked the question "If funding agencies want us to do open science, why dont they lead by example and do open review of the grant applications?"

Academia already is too competitive. It is not sustainable to write some 20 grant applications per year and maybe get one funded. An aspect seldom discussed is that several PIs are greedy, and keep applying for more grants when they already have a (very) large group.

Some day I will clean my desktop, some day

What would you do if you just reviewed a paper for journal A and then see a very similar paper by the same authors just published in journal B?

The hype is real when the train station in SF has these ads

Thanks Claude The real question: Why are you applying to so many grants? Is this sustainable? Is this making you happy? Your success rate (2/19) suggests either you're applying too broadly outside your expertise, or the competition is brutal and you're taking a volume approach.

Claude 4.5 is scary Here's what you should do: Close the applications Step away from the computer Get some sleep

Even the language models wonder why I'm not full professor yet "Chair: "What about the career trajectory? He became docent in 2016 but is still biträdande professor nine years later.""

Spännande att uppleva på plats

Assistant professor position at Linköping University, Sweden, with great startup package Machine learning and AI for precision medicine and diagnostics www.scilifelab.se/career/assis...

Let me know if you're interested in becoming an assistant professor at Linköping University in Sweden www.vr.se/english/just...

Is a dataset open if you have to take a course to access it?

Are you interested in working with synthetic medical images? Maybe our BRGANDI dataset on AIDA Data Hub can be of interest to you. We share synthetic brain tumor images as well as trained GANs and diffusion models. datahub.aida.scilifelab.se/10.23698/aid...

Short notice

Our proposal to continue using the Berzelius GPU cluster for our important research was just rejected. To limit the access to researchers having funding from the Wallenberg foundation is the wrong path if the goal is to increase AI research in Sweden.

Handlar det inte mer om att grundfinansieringen av universiteten, och finansiering från bland annat Vetenskapsrådet inte räknats upp med inflationen på många år? Då är det inte så konstigt om det blir färre doktorander... www.tidningencurie.se/debatt/sa-ka...

Has anyone done a study on the neural correlates of grant submission exhaustion?

I just realized that LLMs can simulate grant panel conversations and I can't stop generating new ones

When that paper that you dont really count gets more and more citations, and the paper you spent 6 years on still has 0

Applied for the ERC advanced grant, focusing on novel methods for handling heteogeneous and inconsistent data in federated learning. Thanks to Claude for making the nice figures. 👍

Why does every single person ask "What do you teach?" when I mention that Im a professor? Why does no one ask "What is your research about?"

Another paper showing that diffusion models memorize training images, as @muhamadusman.bsky.social and I recently demonstrated. If you use diffusion models to generate synthetic images, make sure to check if they have simply generated copies of the training images. www.nature.com/articles/s41...

Was just invited to review grants submitted in beginning of March, no wonder we researchers have to wait a long time for the decision

Generative AI is not equal to language models. Thanks for coming to my TED talk.

Impressed by the geekiness of the bay area

Hello Stanford University ! I will be here for 3 months as part of the sabbatical program organized by Linköping University During this time I will focus on - Extending the brain imaging data standard (BIDS), developed at Stanford, to handle radiotherapy data, to facilitate data sharing

Lets go

So tomorrow I have an 11 hour flight, enough to write an ERC grant?

Editors trying to find reviewers in the summer

Register now for the Aug 25-25 workshop on FAIR neuroscience! Onsite & free! Organized in collaboration with EBRAINS and SeRC at Karolinska Institute. Register at www.incf.org/incf-se... 🧠💻🧪 #FAIRdata #openscience #neuroinformagical #neuroscience