Henning Pohl 📘

New #PhD position in Copenhagen, Denmark 🇩🇰 to do #HCI research on #privacy and #interoperability in messaging apps! Come work with me and @nielsvberkel at @CompSciAAU, fully funded for 3 years. Deadline to apply is May 18. Find me at #CHI2025 to chat! stillinger.aau.dk/phd-stillinger…

A Ph.D. studentship with @jhnhw & myself @UofGlasgow in the ERC DIFAI project will look at developing “mediating mechanisms” as a way to link machine learning to HCI in tightly-coupled control loops. Deadline 14th Jan. findaphd.com/phds/project/p…

@conorlinehan It's pretty ridiculous how we're so strict about stuff like the number of columns in a draft, but have no good process for desk rejecting submissions based on overall quality

Had a bunch of things desk rejected from #chi2025 for the formatting apparently being slightly off, but without any specific feedback about what exactly was wrong. Honestly the formatting should not be a) so difficult or b) deemed so important. waste of everyones time

New article in @acmtochi: We investigate the tensions that the use of large-scale automated edits and bots brings to editors of @openstreetmap. Work together with @henningpohl. Looking forward to presenting this at #chi2025 dl.acm.org/doi/10.1145/36…

Just presented some fun work of five previous students of mine at #imx2024 on how to use generative-AI for personalized media experiences. Paper: dl.acm.org/doi/10.1145/36…, Video: youtube.com/watch?v=P1eBaU…

How should LLMs react if they can’t or don’t want to give you a proper answer? We find that diverting the conversation works surprisingly well for avoiding interaction breakdowns. #CHI2024 @TimSchrills @henningpohl @nielsvberkel Preprint: joelwester.com/wp-content/upl…

For #chi2021, which would have taken place in Yokohama, we had the Hello CHItty plushy. For #chi2025, can we have a piCHIchu?

@ineslrangel Some startup out there is probably already working on an LLM browser extension so things one reads are good vibes only 🙃

@henningpohl Can you write some CSS to make them nicer to read as in nicer things to say about my paper?

Maybe it's useful for others, so here's the CSS I inject into the #chi2024 review page (just another <style> tag in the head) to make them nicer to read (columns, remove header, divider between reviews) when I print them (in landscape format): gist.github.com/henningpohl/d4…

@nbhansen Well, some of RDR2 is quite suiting the CHI submission/review process: youtube.com/watch?v=4gcdoV…

@nbhansen Just answer my doodle already instead of posting about ignoring us here 😄

Hey, @DB_Bahn, this is not a very useful connection your navigator app gives here. Unfortunately, I can't travel back in time when changing trains.

Academic publishers make USD 19 billion+ in sales based on our work. 5 websites you can google to get access to paywalled papers if you can't reach the authors: 1. Sci-Hub 2. Unpaywall 3. 12ft ladder 4. Paperpanda 5. Open Access Button

@nbhansen If you need an agitated German, I can get you an agitated German. No problem.

@dblp_org @acm_chi @ACMDL @sigchi I've had some papers missing, author orders mixed up, and incomplete data (mostly dois missing). Already sent in the ones I encountered, but makes me wary overall. That said, that data also has some issues right at the source (ACM in this case), so I wouldn't blame dblp for that.

@henningpohl @acm_chi @ACMDL @sigchi May I ask what kind of issues you are having with the #dblp metadata? ~MRA

Trying to run some automated analyses of @acm_chi papers and the @ACMDL and @sigchi metadata is often such a mess. Differing acceptance rate info (see e.g., CHI '15), wrong paper types, numbers off, chair info wrong, ... Tried data from @dblp_org, which also has some issues 🤷‍♂️😟

Attention👋! We are looking to further expand our academic capacity with a number of Associate and Assistant Professors (+TT) in #computerscience More information👉cs.aau.dk/Vacancies/ #hiring #jobsearch #recruiting #applynow #sciencejobs

I guess there are some uses for the Crossmark button (dl.acm.org/about/crossmar…) now apparently added to pdfs on the @ACMDL, but why is this a printable element? Did nobody think this through?

It's just 6 datasets from me & colleagues right now and thus also not very diverse. If others are happy to donate some of their own data, that would be great! No personal data + all coauthors need to be on board. Just submit a pull request on github: github.com/henningpohl/hc…

When teaching quant. methods, I found it quite annoying how there is a lack of #hci dataset that are easy to use during class and in assignments. To fix that, I've started the hcidata package for R, now available on CRAN: cran.r-project.org/web/packages/h…