Going to a ML conference from the comfort of my own city is a special and exciting thing happening next week at @euripsconf.bsky.social in Copenhagen.
Looking forward to give our latest ICML paper another presentation at ELLIS UnConference and see how our new @neuripsconf.bsky.social paper fares!
If you want to meet up at any point to talk about
-> physics-informed ML, autodiff, linear algebra, Bayesian DL, or probabilistic numerics,
just let me know. Always happy to talk research, collaboration, (jobs!), or just grab a coffee/beer.
Looking forward to seeing everyone in Copenhagen! 🇩🇰
Saturday/Sunday: I will likely attend the Differentiable Systems [5] (a topic dear to me) and PriGM [6] workshops, which look to have a neat line-up of presenters/papers. Looking forward to that!
[5] differentiable-systems.github.io/workshop-eur...
[6] sites.google.com/view/prigm-e...
Wednesday: We'll be presenting VIKING, which does variational inference for deep neural networks by leveraging some recent developments in Laplace approxs [4].
Lead author Samuel is in San Diego, but several of us will be in Copenhagen and very ready to chat.
[4] Paper: arxiv.org/abs/2510.23684
Before that, I will likely be at the amortized ProbML workshop [3], which looks really nice.
[1] Paper: openreview.net/forum?id=LVQ...
[2] Paper: proceedings.mlr.press/v271/kramer2...
[3] Workshop: sites.google.com/view/amortiz...
Tuesday: I present a poster on numerically robust fixed-point smoothing [1] (TMLR).
Fixed-point smoothing = conditioning on future observations in (linear) Gaussian state-space models.
New algo: both fast *and* reliable, useful for (eg) probnum ODE solvers [2].
Happy to discuss at the poster!
Really looking forward to Eurips next week! 🇩🇰 😎
Not just because Bella Centre is literally a metro ride from home, but because it's really excellent to see the ML community gathering in Europe!
If you're around, say hi! Here's where I'll be 🔽🔽
Accepted papers at ProbNum 2025 have now been published as a volume in Proceedings of Machine Learning Research (PMLR). Thank you to all the authors for their beautiful papers and to the PMLR editor Prof Neil Lawrence @lawrennd.bsky.social for making this possible.
proceedings.mlr.press/v271/
EurIPS includes a call for both Workshops and Affinity Workshops!
We look forward to making #EurIPS a diverse and inclusive event with you.
The submission deadlines are August 22nd, AoE.
More information at:
eurips.cc/call-for-wor...
eurips.cc/call-for-aff...
🌦️ A generative diffusion model that maps coarse climate simulations to a posterior distribution over coherent local weather patterns.
I’d be curious to hear what you think!
🔗 Paper: www.nature.com/articles/s41...
Very nice!
1/
🚨 New paper at #ICML2025!
Identifying Latent Metric Structures in Deep Latent Variable Models 🎉
We solve part of the identifiability puzzle in generative models — using geometry. 🧵
Nicholas Kr\"amer, Filip Tronarp
Numerically robust Gaussian state estimation with singular observation noise
https://arxiv.org/abs/2503.10279
Awesome, thanks!
Looks great! Unless I've missed something, the style file does not contain any information about appendices ("appendix" vs "supplement", same/separate file, single/double column, etc.). Is that intended?
Big news: you can now submit your papers to the first ever international conference on probabilistic numerics (1-3 September, Southern France)! Don’t miss the chance to get in on the ground floor of this exciting new field! probnum25.github.io/submissions
Numerically Robust Fixed-Point Smoothing Without State Augmentation
Nicholas Krämer
Action editor: Jake C. Snell
https://openreview.net/forum?id=LVQ8BEL5n3
#smoothing #robustness #robust
This is really cool!
Poster session happening *today* at 4:30 local time.
*East* Exhibit Hall. Poster #3511.
Looking forward to presenting this work! See you there? 🙂
Heading to Vancouver for NeurIPS to present our paper “On Conditional Diffusion Models for PDE Simulation”. I'll be together with Sasha and Cristiana at poster 2500 during Thursday’s late afternoon session. Looking forward exciting discussions and meeting new people! 🥸🥸
neurips.cc/virtual/2024...
📢 On Sunday, catch me at the D3S3 workshop (d3s3workshop.github.io), where I'll talk about probabilistic numerics and simulation from 1:00–1:30 p.m.
This workshop gonna be great fun!
Looking forward to connecting with everyone!
This is not just for the autodiff and linear algebra folks. The paper also discusses PDEs, Gaussian processes, and Bayesian neural networks. Oh, and there is also plenty of #JAX code: github.com/pnkraemer/ma...
⭐ On Wednesday, I'll be presenting a poster on our spotlight paper:
Gradients of Functions of Large Matrices (arxiv.org/abs/2405.17277). Stop by Poster Session 2 East, Wed 11 Dec, 4:30–7:30 p.m. to say hi! :)
🛩️ On my way to #NeurIPS2024 and excited to chat about (ML applications of) linear algebra, differentiable programming, and probabilistic numerics!
Feel free to DM if you’d like to meet up, hang out, and/or discuss any of these topics 😊
(Where to find me & paper info? -> Thread)
I don't think I can help with that, sorry.
This is gonna be awesome!
We'll be around to discuss evaluating gradients of functions of la(aaaa)rge matrices arxiv.org/abs/2405.17277 and possible applications in GPs, PDEs, BNNs and beyond 😊
@nathanaelbosch.bsky.social
Really looking forward to this! 🤩
Trembling with excitement in announcing the first ever International Conference on Probabilistic Numerics! Probabilistic Numerics is about machine learning FOR numerical computation, like optimisation. 1st - 3rd September 2025, EURECOM, France, submissions due 5th March 2025. probnum25.github.io
Elizabeth L. Baker, Moritz Schauer, Stefan Sommer
Score matching for bridges without time-reversals
https://arxiv.org/abs/2407.15455
