For a recent lab meeting, I wrote up a grab bag of ways to think about your development as a researcher during a PhD: emerge-lab.github.io/papers/an-un...
Sharing in case folks find it useful or have feedback!
Flash-KMeans
An IO-aware implementation of exact k-means that rethinks the algorithm around modern GPU bottlenecks. Flash-KMeans achieves 30x speedup over cuML and 200x speedup over FAISS.
Paper: arxiv.org/abs/2603.09229
Code: github.com/svg-project/...
Claude Code Copenhagen meetup 🍻🚀
First blogpost of my cofounder post-Delhi on the deafening silence of public/political sphere as we come closer to another industrial revolution. anastasiastasenko.substack.com/p/the-countr...
Strongly agree! Sounds like a fantastic idea for a special course :) Maybe you can take inspiration from Danoliterate danoliterate.compute.dtu.dk which was a master's project at DTU
Helt klart. Der er stadig tonsvis af quirks. Jeg er mest begejstret for tendensen der bliver vist. At software og hardware begge udvikler sig i retning af at lokal inference bliver konkurrencedygtigt ift. cloud.
LLM cloud inference dominates usage, but should it? Local models and accelerators have improved massively over recent years.
Perfect routing to best local model "reduce energy consumption by 80.4%, compute by 77.3%, and cost by 73.8% versus cloud-only deployment"
arxiv.org/pdf/2511.07885
Introducing ✨Tiny Aya✨, a family of massively multilingual small language models built to run where people actually are.
Tiny Aya delivers strong multilingual performance in 70+ global languages in a 3.35B parameter model, efficient enough to run locally, even on a phone.
Excited to be in fantastic company amongst the speakers for IDA Driving AI 2026 :)
ida.dk/driving-ai/t...
The paper suggests that poor reasoning abilities is one failure mode that comes from removing deeper layers. But yeah, wonder what else unexpectedly takes a hit
Due to residual structures in Transformer-models it's possible that many layers contribute very little to the downstream performance of the network, allowing for removal those redundant layers with little impact.
openreview.net/pdf/72eeca23...
Local AI inference through the browser is great. Transformers.js makes it feasible for someone like me who has little web dev experience :)
Here's a simple transcription tool using Whisper-large-v3-turbo, making use of your local hardware and ensuring privacy.
rasgaard.com/webai-stuff/...
ofc there's already a python plotting library for this sort of thing
github.com/jndean/LossR...
This is crazy:
youtu.be/v81b0XllvgI
www.youtube.com/watch?v=b64v...
A photograph of sunny Copenhagen in the summer!
📢 I am hiring a highly-motivated Ph.D student at the University of Copenhagen to work on tokenization-free NLP.
Read our previous work in this topic: aclanthology.org/2025.emnlp-m...
aclanthology.org/2023.emnlp-m...
openreview.net/forum?id=FkS...
Apply by March 8: employment.ku.dk/phd/?show=1563
Sometimes a single line changes the reproducibility-game completely:))
The paper is really cool though: Can the transformer+conv Mimi (NAC) decoder be replaced with a purely transformer-based one?
Answer is yes, and it's 10x faster! Benchmarked on actual mobile hardware.
arxiv.org/pdf/2601.20094
WE ARE BACK! 🚀🚀🚀
📢 #databeers #copenhagen x Absalon is back on March 10th, SAVE THE DATE!
Many thanks to our sponsors, the Danish Data Science Academy and the Danmarks Tekniske Universitet
Speaker announcement and tickets out soon, stay tuned!
Slowly but surely getting a better intuition for how and when compression methods such as quantization is beneficial and, perhaps more crucially, detrimental to overall model efficiency
"numerical precision reduction yields the most benefit in the prefill phase of large models, where compute dominates. In contrast, the decode phase remains memory-limited, and aggressive quantization (e.g., int8 or int4) may incur overheads that outweigh theoretical savings"
It took me weeks, but finally it's there: an overlong blogpost on synthetic pretraining. vintagedata.org/blog/posts/s...
"Når 32 % af alle danske skoler bruger OS2 Skole, går det i nul. Omkostningerne ved OS2 Skole vil ikke være de licenser, der betales til Big Tech, men penge, der betales til de [skattebetalende] personer, der ansættes til at vedligeholde og udvikle systemet."
dataethics.eu/da/new-open-...
This reddit post made me laugh
Kan man klare sig uden amerikansk 'big tech'? Her er nogle af alternativerne www.dr.dk/nyheder/viden/teknologi/...
#dk #nyheder #dkbot
In Grenoble/France where I work, a steep rise in bicycle commutes (green) happened, mainly because of strong support from the local municipality through infrastructure changes - bike lanes! This coincides with a drop in car usage (red), while walking (yellow) and public transp. (black) stayed flat.
The winners were,
- CoRal: huggingface.co/CoRal-project/
- Top Scandinavian text embedding model: huggingface.co/jealk/TTC-L2...
- EuroEval: euroeval.com
Had a great time at @danskerhverv.dk today where the Danish Data Science Community (ddsc.io - join the Slack!) revealed the winners of the first Open Source Awards
