2025 in Numbers

Tags: news, musings

Published on
« Previous post: AI: From Homer to ChatGPT

Continuing the tradition, here is a brief, number-based summary of 2025.

As usual, the good old academic KPIs come first since, in the absence of knowing what is important to measure, we settle for measuring that which is convenient. In 2025, the AIDOS Lab

  • …published 12 conference or journal papers (with 7 of them appearing in my favorite machine learning conferences)
  • …published 5 workshop papers
  • …wrote 11 preprints

The lab also grew quite a bit, with 2 postdocs and 4 doctoral students joining. This is wild! 🤯 (And most of the funding comes from my ERC Starting Grant, graciously funded by the Swiss State Secretariat for Education, Research and Innovation.)

We had two short-term academic visitors, i.e., Victor and Giacomo, and one long-term visiting professor, Nello, who chose to spend his sabbatical here (and will stay for a couple of months in 2026 as well).

This time, I wrote 192,023 words in my primary documents repository. What difference a year not on the job market makes! Most of my writing was done outside this repository, i.e., in e-mails, papers, presentations, or grants, all of which are tracked differently.

Speaking of presentations, I prepared 19 talks for a variety of audiences, my favorite one being a talk on modern AI, which I gave in French to simplify matters for an audience hailing primarily from Romandy. To borrow a phrase from the inimitable Eamonn Keogh: The talk went well and was well-received.

In terms of acceptances and rejections, I got 12 papers or grants rejected, but 20 submissions worked out in my favor (given the glacial speed of academia, this does not necessarily imply that they appear in early 2026, though). These are fewer rejections than in 2024 and 2023, which I would generally take as a good sign: I am still trying out lots of things but I am getting better at them (or maybe I am just on top of the Wheel of Fortune for a while now).

I could write about even weirder academic KPIs but I want to rid myself of the need to put stuff that ultimately does not matter too much into numbers. Suffice it to say that the work is still having traction and I am enjoying my collaborations immensely.

Let’s make 2026 a great year, everyone!