On Mastodon I’ve been enjoying the #3GoodThings hashtag where people post about good things of that day (or perhaps a longer time period). I just posted one of my own there, and then I thought in general this concept might be a good way to carefully start again.

It’s 8AM here so not a lot has happened today, but here are some good things I wanted to share.

Thing #1

We submitted a paper with Amelia, several other people in the lab, and new collaborators we met only recently. This was what I like to call a 1+1=3 collaboration. To me these are rare but extremely valuable, and I hope we will be doing more of this.

In general recently several other collaborations with lab members and others have been quite inspiring, so look out for more #3GoodThings about this πŸ™‚

Thing #2

A colleague wrote me a “happy folder” email about this blog. I realize I’ve been neglecting this space for a while, but this is still so good to hear. Together with a conversation with another colleague about blogs, and the #3GoodThings motivation, maybe this is how I will get to do more of it again.

Thing #3

While working on the paper from thing #1, we stayed late at the office and got pizza. Amelia found a pizzeria close by, but they did not do delivery and I thought I could use the fresh air. The amazing things that happened after is that I actually (i) called them on the phone and (ii) spoke Danish for the entire process. I was a bit less impressed with myself when I realized the person I spoke to was probably Italian, but I’ll still take it as a win.

//Perhaps it’s important to note here that this type of event is quite a rare. I usually leave the office before 5PM and discourage working overtime with people in the lab. But once in a while, it can be a good experience πŸ™‚


These were the 3 things! Maybe, I will post something else soon again

Cats in my research

I have had the idea to post this for a while, but now it’s finally happening! Somebody on Mastodon is going to write a review of cats appearing in papers, so the best I can do is provide references πŸ™‚ Here are some ways I added cats to my papers, in chronological order.

2013 – Cat mention

Three years into my PhD is the first mention of cats in my papers. In “Combining Instance Information to Classify Bags” (PDF here) which is about my PhD topic of multiple instance learning, I write “For example, an image labeled as β€œcat” would have a cat in at least one of its segments, whereas images without this label would not portray any cats at all“.

I’m not sure whether this was a coincidence or intentional. But more cats happened in the years after!


In 2015 I finished my PhD thesis, and the cover of the thesis features a puzzle with cat illustrations! The puzzle could have worked with any kind of illustration, so the cats were definitely intentional. Thanks to Hella Hekkelman who made the illustration.

Unfolded book cover titled "Dissimilarity-Based Multiple Instance Learning" by Veronika Cheplygina, both front and back covers have groups of cartoon cats on them.

I also added my cat Buffy to the thesis acknowledgments!


Probably the biggest achievement of my career – I wrote a paped titled “Cats or CAT scans: transfer learning from medical or natural image source datasets” (PDF here) AND inserted a picture of my cat Buffy (by then over the rainbow) into it. Here is Buffy pretending to be an image from ImageNet:


More images of Buffy, now occupying the whole figure, in “Ten simple rules for getting started on Twitter as an academic“. Now teaching the reader about cat-related hashtags like #AcademicsWithCats and #Caturday.


Based on the work about transfer learning I stared a few years before, I received a grant from the Novo Nordisk Foundation called “CATS: Choosing A Transfer Source for medical image classification”.


More Buffy, promoting self-care in “Ten simple rules for failing successfully in academia“.

More cats?

That’s it as far as I can remember, but Pixel and Dot definitely deserve paper appearances as well, so keep an eye out for that πŸ™‚

Postdoc position on transfer learning and dataset similarity

I’m hiring a postdoc position, starting 1 September 2023 or shortly thereafter, until 31 August 2025. You can find the full vacancy text on the ITU website.

The project is titled “CATS: Choosing A Transfer Source for medical image classification” and aims to investigate the different ways to assess dataset similarity in relation to that dataset’s “transferability” to a medical target task. This preprint by Dovile already has some recent results from this project.

Please forward to anyone interested in doing research with us in Denmark! To apply, please go through the HR system above, or get in touch via vech AT itu.dk if you have any questions.

Research update – lab website, papers & webinar

A lot has been happening in research life the last year! In this post I’m sharing a couple of recent updates.

Lab website

In 2022 three people – Bethany, Dovile in January, and Amelia in August, joined the lab. Although I used the name “PURRlab” already when I was in Eindhoven, this time we went a bit further and made our own lab website and everything.


The latest news is that the paper on shortcuts in chest X-rays (see preprint), led by Amelia, has been accepted at ISBI 2023! Here’s a tweetorial by Amelia about the work.

There are more preprints coming up in the next few months on various topics related to datasets, transfer learning, bias, and understanding ML researchers.


We are also organizing a webinar series where we want to take a deeper look into the datasets used in machine learning for healthcare. In the first edition we have three speakers who will talk about their work on skin lesions datasets, but we will be exploring other applications in future editions. If you want to stay updated, we have a newsletter you can sign up for on the webinar page. Thanks to Eike Petersen (DTU) for help with the setup!


I’m on Mastodon, and I love it

Things are happening quickly on the bird site! Time for a blog.

When the possible acquisition was announced back in April, I created my Mastodon account at scholar.social. (If you haven’t heard of Mastodon, there are various great guides, for example here and here). Back in April it was a bit quiet and I didn’t use it much, but things feel different this time.

My bird account dates back to 2010, but I only really started using it in 2015 or so, as a postdoc. I was struggling in various ways during that time (not a secret, but not the point now, if you are curious you can infer a few things from my shadow CV). But with #AcademicTwitter, I found a community, and it had a lot of impact on me as a person. Being there led to many good things, and I even wrote a paper and gave some talks about it.

Twitter changed a lot since then. I did my best to keep my experience algorithm-free, by using a chronological timeline, muting suggested tweets and such (example). I like to think I was quite successful in this, because my Twitter experience has been good. But the recent activity on Mastodon made me realize that recently my Twitter experience wasn’t great.

I don’t want to get into too many details right now on why I think Twitter is not a good place to stay or trying to convince others about this. Perhaps it’s worth mentioning I hard quit various social apps a few years ago, due to both mental health and ethical considerations, and more recently I’ve replaced many big tech services with open source alternatives. All of this has been great, so if it seems surprising that I’m not so lyrical about Twitter right now, it’s actually not.

With my reasons, a few weeks ago I dusted off my account on scholar.social, and I absolutely love it. I found things that I remember from “early Twitter”, like connecting about niche topics. There are no ads, no algorithmic suggestions, and lots of helpful people. There is a feeling of excitement, like we are all going on a road trip. I genuinely missed this feeling, and didn’t even realize it – until the recent events became the catalyst. I’d like this feeling to stay, and I want to contribute to building the community. I did already move instances to dair-community.social, as I’ve excited to see their local timeline. So from now on you will mainly find me @DrVeronikaCH@dair-community.social, and of course, this blog πŸ™‚

10 things I miss about The Netherlands

Later this week it will be 1.5 years since I moved to Denmark, and a Twitter poll told me that you would like to hear what I miss about The Netherlands (NL)! It goes without saying that of course I miss many people there who I don’t get to see as often! But here are 10 things about NL that perhaps I was not expecting to miss.

1 – Healthcare professionals who listen to me

I’ve had to deal with a number of both physical and mental health issues in recent years, and in NL I finally felt like I was getting the care that I needed. To put this mildly, the level of ignorance both me and my husband encountered here would be the number one reason for us to go back, and to advise anyone in similar situations to stay away. I don’t want to go into all the details, but perhaps the best example is my psychiatrist saying “I don’t remember what dose you are on but just double it” about lithium, after I had described the adverse side effects I’ve been experiencing.

2 – The train connections

I like travelling by train and although Denmark has good connections to Sweden and Germany, it was easier to reach more places directly from NL – London in 4 hours, Paris in 3, Berlin in 6 and so on.

3 – The tap water

Tap water in Denmark is safe to drink, but has high concentrations of minerals (safe but annoying to deal with due to the limescale build-up) and the tap water in NL just somehow tastes better to me.

4 – My go-to shops

In NL I developed a list of places that I’d buy a lot of basics, that I haven’t quite found good replacements for here. My top three are (the links go to Wikipedia for some more information about the stores):

  • Decathlon, which sells sports and hiking equipment, mostly of in-house brands which have great designs but are a lot cheaper. Decathlon exists in 60 countries, but not in Denmark, and none of the neighbouring countries want to ship here.
  • Hema, which sells pretty much everything, again with own designs and great prices. It is a Dutch store that has a few stores in other countries. I haven’t tried shipping yet, but there is also something nice in just browsing there in person.
  • Amazing Oriental, a grocery store with food and drinks from many Asian cuisines. I’ve found smaller stores here in Copenhagen, but none that have quite the same range of products in one place.

5 – The freshest vegetables

NL produces a lot of vegetables, and they stay good longer, compared to some vegetables that Denmark imports from Spain, or, yes, The Netherlands. Or maybe it’s just the grocery store next to my place. I don’t miss a lot of specific foods otherwise, maybe except proper pindakaas (“peanut cheese”).

6 – Sauna complexes

You would think that the proximity to other Nordic countries would mean this is a thing here. Yes, there are saunas, but mostly as a “side” to swimming, and/or by the hour, and/or are “not a real sauna” according to my Swedish husband. I’m talking about complexes where you buy a day pass and can go to multiple saunas, hot baths, cold baths, a restaurant, and so on – popular in NL and I think at least in a part of Germany.

7 – At least 5K EUR extra that used to be in my bank account

Although I received some moving allowance from ITU, it just covered the cost of the movers driving here with our things, and we did all of the packing etc ourselves to reduce costs. But with a few trips back to NL with certified PCR tests at each end, not being to go across country borders when originally planned, and ultimately furnishing an entire apartment in a few days, it was even a more expensive & stressful exercise than moving countries is otherwise. To be fair, this was not entirely unexpected, and one of the reasons I never understood how some senior academics expect juniors to just go and do that multiple times (the other is healthcare, as per point 1).

8 – DIY-ing the place I live

Alterations are not allowed in many? rentals, therefore also not where I live now. I miss being able to attach things to walls, ceilings, and other existing structures. First thing I’m doing when this is possible again, is getting one of those ceiling bridges for the cats.

9 – A scheduled lunch break

Lunch food in Denmark is much better – I was never quite satisfied with NL’s selection of bread with cheese and milk (yes, really). But at least in Dutch universities there is a lunch break, say from 12:30 to 13:30, when no classes are scheduled. If a “lunch meeting” is scheduled, usually said meeting will at least have the bread/cheese/milk on offer. Here classes start at even hour blocks (8am, 10am, 12pm, 2pm, 4pm) and it can be difficult to meet with colleagues who are teaching or have regular meetings without food at 12pm.

10 – Oranjekoorts a.k.a. “Orange fever”

A thing that happens during King’s (previously Queen’s) day, and whenever NL plays football where everybody dresses up in orange and paints Dutch flags on their face and it’s a big party, like so:

A woman and a man in their thirties, wearing orange clothes, red/white/blue flags on their faces and red/white/blue flower guirlands.

Postdoc position in machine learning & image analysis

Last year next to the NovoNordisk starting grant which funded two PhD positions, I was also lucky enough to be awarded the Inge Lehmann grant from the Independent Research Fund of Denmark, where I will be hiring a postdoc!

The project is titled “Making Metadata Count” and aims to investigate the role of metadata in machine learning methods for medical images. This includes both “traditional” metadata like patient age, and other image information, like the presence surgical markers. This is a 3 year position starting October 2022 (negotiable), and the call closes on the 30th of April – for full details please see the vacancy on ITU’s website.

Please forward to anyone interested in doing research with me in Denmark! To apply, please go through the HR system above, or get in touch via vech AT itu.dk if you have any questions.

7 things I wish I would have done during my tenure track

Recently I’ve seen some Twitter threads on advice for new PhDs/postdocs/PIs. I’ve shared this about my PhD before (see 7 things I wish I would have done during my PhD), and in my current job I’ve been reflecting a bit about how my previous job went, I thought I would also share the 7 things I wish I would have done during my tenure track!

They are:

  1. Being in the office less
  2. Not agreeing to only teach undergraduate courses
  3. Spending less energy on grants
  4. Divorcing my email accounts
  5. Getting a Macbook
  6. More papers with people from Twitter
  7. Sharing more work online

1 Being in the office less

After starting my tenure track in February 2017, my best-case scenario day would look like this: I would leave my house in Rotterdam at 7:20, to take the train and get to the office at 9:00. Similarly I would leave at 16:50 and get home at 18:30, in time for dinner. There were sometimes disruptions, with me arriving 1-2 hours late in either direction. Maybe not too bad considering depending from where in the world you are reading. And as a plus I could work in the train – I would often read or draft blog posts.

After two years of working at home, I cannot imagine being able to do these kind of hours again. Although I was already quite mindful of the hours I would spend on my job, I didn’t realize that doing productive things in the train also counted, and that I probably was not getting enough rest. I also don’t really understand why I felt it was necessary, as I wasn’t required to be in the office on specific days or hours unless I was teaching or had other meetings.

2 Not agreeing to only teach undergraduate courses

For the first three years, I was the course manager of a first year BSc project and taught in another third year BSc course. I enjoy teaching at multiple levels, but I think I shot myself in the foot a bit here. As I had no start-up, PhD students to co-supervise, etc, the main way to start new projects was to supervise MSc students. But since I wasn’t a person the then-MSc students were aware of, recruiting such students was rather difficult.

3 Spending less energy on grants

Funny given that during my PhD I wish I had “applied to all the things”! As part of my tenure track conditions I had to apply for two “medium” (1 PhD position size) grants a year, which was reasonable, and it was useful to think about the project proposal. But several of these applications were doomed to fail, as even with a perfect score on the research proposal, my CV just was not “good enough” to get funding. Given that there weren’t many other opportunities to apply for, I guess overall this was still a useful experience, but I definitely could have spent less time on writing workshops, endless revisions etc.

Another advice was to apply for all possible small grants (workshops, collaborations) that I could get. I did that and actually got several of the things I applied for. But this was too much – relative to larger proposals, these cost more time to write, AND require more work from you after, without the option of hiring somebody to help.

4 Divorcing my email accounts

I used be very much a “one inbox” kind of person, and forwarded my university email to my Gmail. But with a new job and email account, I decided to try it out. Although Gmail has a much better interface than Outlook (don’t get me started on this…), I like it a lot. I don’t have Outlook on my phone, so I mainly have access to my work email during my work hours. This frees up a bit of headspace during time off, which I would often already use to mentally draft emails, thus spending way more energy on emails overall.

5. Getting a Macbook

Similarly to me trying out a different strategy with emails, I felt brave enough to try out a Macbook after a lifetime of Windows. It’s kind of great, I’m still pretty inept at using shortcuts etc, but I don’t imagine going back anytime soon.

6. Starting more papers with people from Twitter

One of the most satisfying things in my career has been to work with people met on Twitter. The prime example is probably this paper about Twitter – some of us have met each other beforehand but I feel safe to say this was a Twitter collaboration.

Another highlight was this preprint with GaΓ«l – it started because I agreed with him about AI being hyped too much and said “we should write a paper about this”. A longer revised version has just been accepted so keep an eye out.

I like slow science, so there are other projects that are amazing and that started on Twitter but are not out yet. Stay tuned πŸ™‚

7.Sharing more work online

This perhaps sounds surprising since I share preprints, slides etc and would even write blog posts on a more-regular basis. But I still see so many things that I’ve done that could be possibly useful to others, that I did not share, either because the thing needed a bit more input (for example going from an undergraduate project to a preprint), or simply because I didn’t get around to it (for example rejected grant proposals).

I am not sure I will ever get to a point where I’m doing this better, but as usual, if there is something I have that might be helpful to you, just ask.


This is just my list! There are also a few issues others mentioned in the Twitter discussion that maybe I did OK with? and are therefore not on this page – but that’s for another blog post πŸ™‚

Rethinking productivity

Hello friends,

It goes without saying that I haven’t been on this website much lately. You would maybe expect that the explanation would be that “I’m busy” because I got a new job and moved countries. But while it’s certainly been a moving year (pun intended), that’s not quite it. I had time to blog. I just didn’t want to.

This goes against what I used to be recommend to others from this very website, such as doing a little bit each day. I still agree that’s a good way to get started with a project and to get things done. But I realized it’s not the best strategy for me on a larger scale of things. Even with all my productivity ideas (coping strategies?), my brain gets tired of doing multiple projects at the same time. And with everything happening this year, there was time – but no space – left for other things. But things are slowly converging, which is why I’m writing this πŸ™‚

In general I think quite a few things changed in how I approach productivity compared to a few years ago. I tweeted a bit about this recently, and as you can see, the thought process is still ongoing. But here are things that I’m leaning more towards now, or some things that didn’t work quite as expected.

Having a few focus projects

I need to have only a few things I’m doing on a given day or week. I already liked the idea, as described in my Kanban post. But I try to take it into account more now, and I think I’ve been mildly successful with not starting new projects before I finish existing ones.

In a way, this also means I cannot keep up with all the daily habits I would like to have. If I start writing the first thing in the morning, I will probably end up writing for hours, and maybe forget to eat (and definitely “forget” to exercise). On the positive side, I get more done in that type of day, than if you would just take the hours put together, so I think I’m just going to run with it.

No work email on my phone

I was a big supporter of the “one inbox” principle, but ten years into academia, I decided to give it a try when I started at ITU, and finally use Outlook without forwarding mail to my Gmail. The headspace this gives me on my days off is excellent.

The disadvantages are that Outlook is not great with search, and although there is integration with Todoist, the URLs that are created out of tasks often fail. But if I cannot 100% decide how I want to do email, this will have to do.

Scheduling tasks on the calendar

I already tried to do this with Getting Things Done but I think that level was too fine-grained, and difficult to keep up with. Since I also leveled up to actually using my work’s Outlook calendar, I reserve 3-6 hour blocks to finish tasks which are important and coming up soon, such as preparing talks or grading exams. It often ends up being less hours and ends up getting moved around, but it helps me see how much I have to do in a particular time period.

(Not) capturing everything

Again in the Getting Things Done category, I was happy to have a system to capture everything I might need to think of later. I have relaxed this a bit, some things don’t need to be captured – at least by me. I am still saving a lot of papers I probably won’t read in full, but otherwise I’m not using Evernote quite as much as I used to.

Except Fitbit for health purposes, I’m also not using any previously-tried “productivity apps” that track how long I am using which app, how long I am using my phone, etc.


That’s all I have at this point, but I’m planning to come back a bit more often, so let me know if there’s something specific you’d like to hear more about. For now, I’ll get back to my focus project of organizing my apartment πŸ™‚

Two PhD positions in medical imaging

I’m happy to announce that NovoNordisk Fonden recently funded by project “CATS: Choosing a Transfer Source for medical image classification”, which means I will be hiring two PhD students in the coming months!

The project is inspired by my earlier paper “Cats or CAT scans: Transfer learning from natural or medical image source data sets?” [also on arXiV] where I showed that there is no consensus on how to select a source dataset to initialize weights of a model, that will be further trained on medical imaging.

The project will focus on investigating dataset similarities, which will help us define how to select a good source dataset, and PhDs 1 & 2 will investigate different approaches to defining this similarity.

I am looking for two PhD students with different backgrounds (math/computer science, and psychology/computational social sciences) for the project. You can view the vacancies here (closing date July 31st):

  • PhD 1 (maths/computer science)
  • PhD 2 (psychology/social sciences)

Please forward to any interested candidates! To apply, please go through the HR system above, but I am happy to answer questions about the project, just drop me an email πŸ™‚

Mastodon More Mastodon