Motonobu Kanagawa's Avatar

Motonobu Kanagawa

@motonobu-kanagawa

Statistics, machine learning, numerics, and simulation at EURECOM.

2,217
Followers
620
Following
69
Posts
17.11.2024
Joined
Posts Following

Latest posts by Motonobu Kanagawa @motonobu-kanagawa

Please feel free to share with anyone who may be interested. The registration deadline is February 13th.

29.01.2026 09:12 πŸ‘ 0 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0
Scientific Program – NUMRAD26

In June, I will give a lecture on ``Gaussian processes and reproducing kernels'' at the NUMRAD26 summer school held at EPFL.

numrad.epfl.ch/scientific-p...

This will be an exciting event for math-oriented students interested in numerical methods, statistics, and machine learning.

29.01.2026 09:12 πŸ‘ 5 πŸ” 0 πŸ’¬ 1 πŸ“Œ 0
Preview
New method improves the reliability of statistical estimations MIT researchers developed a method that generates more accurate uncertainty measures for certain types of estimation. This could help improve the reliability of data analyses in areas like economics, ...

I’m excited that MIT News covered our new paper on confidence intervals for associations in spatial settings!
news.mit.edu/2025/new-met...

12.12.2025 19:12 πŸ‘ 23 πŸ” 7 πŸ’¬ 2 πŸ“Œ 0

Convolutional Fenchel-Young loss paper has been updated slightly, with more examples e.g. classification with rejection (with the great effort of
@caoyuzhou.bsky.social
). We're gonna present the spotlight poster at #NeurIPS2025, at 4:30pm-7:30pm on Dec 3. See you all soon😎

29.11.2025 06:11 πŸ‘ 6 πŸ” 2 πŸ’¬ 1 πŸ“Œ 1

Really looking forward to Eurips next week! πŸ‡©πŸ‡° 😎

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 πŸ”½πŸ”½

27.11.2025 11:00 πŸ‘ 11 πŸ” 5 πŸ’¬ 1 πŸ“Œ 0
Post image

Our paper on missing-value imputation has just been published in Transactions of the Machine Learning Research (TMLR)! Our kNNSampler recovers missing-value distributions with theoretical guarantees, while being as simple as kNNImputer.

openreview.net/forum?id=4CD...

02.12.2025 08:57 πŸ‘ 4 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

Big thanks to @natematias.bsky.social for this extremely touching and kind post about some of our research! (Amazing co-first authors are my PhD student @renberlinghieri.bsky.social and postdoc David Burt, and awesome collaborators are Paolo Giani and Arlene Fiore.)

26.11.2025 17:16 πŸ‘ 7 πŸ” 3 πŸ’¬ 0 πŸ“Œ 0
Post image

Find out more at probnum2026.github.io

The deadline for Submissions will likely be in early March 2026.

12.09.2025 07:56 πŸ‘ 4 πŸ” 1 πŸ’¬ 0 πŸ“Œ 0

f you're looking for a place to share your ideas on computation and inference, with likeminded, friendly colleagues who give earnest, technically deep feedback, mark your calendars for 9-11 September 2026, because next year, we're going to Finnland, thanks to Toni Karvonen!

12.09.2025 07:56 πŸ‘ 6 πŸ” 1 πŸ’¬ 1 πŸ“Œ 0
Post image

The very first international conference on Probabilistic Numerics is now over. Huge thanks to @motonobu-kanagawa.bsky.social for hosting us at the French Riviera!

It was amazing to see so many contributed works across Europe. Proceedings at proceedings.mlr.press/v271/.

Read on for what's next!

12.09.2025 07:56 πŸ‘ 15 πŸ” 3 πŸ’¬ 1 πŸ“Œ 0
Post image

Our new paper on kNNSampler fixes the widely used kNNImputer for missing value imputation, to recover the distribution of unknown missing values.

Theoretical convergence guarantees in terms of conditional RKHS embeddings.

arxiv.org/abs/2509.08366

11.09.2025 09:43 πŸ‘ 7 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

Parastoo Pashmchi, Jerome Benoit, Motonobu Kanagawa
kNNSampler: Stochastic Imputations for Recovering Missing Value Distributions
https://arxiv.org/abs/2509.08366

11.09.2025 04:08 πŸ‘ 1 πŸ” 1 πŸ’¬ 0 πŸ“Œ 0
Post image

Quite funny:

08.09.2025 06:59 πŸ‘ 20 πŸ” 4 πŸ’¬ 1 πŸ“Œ 0
Post image

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/

07.09.2025 15:03 πŸ‘ 14 πŸ” 5 πŸ’¬ 0 πŸ“Œ 0
Post image Post image

Thank you very much @arthurgretton.bsky.social for your great talk!

04.09.2025 13:14 πŸ‘ 3 πŸ” 0 πŸ’¬ 0 πŸ“Œ 1
Post image

ProbNum 2025 Keynote 2 ``Gradient Flows on the Maximum Mean Discrepancy'' by @arthurgretton.bsky.social ( @gatsbyucl.bsky.social and Google DeepMind.

Slides available here: probnum25.github.io/keynotes

04.09.2025 13:07 πŸ‘ 6 πŸ” 2 πŸ’¬ 1 πŸ“Œ 0
Post image

Getting very much ready for ProbNum 2025!

probnum25.github.io

29.08.2025 13:24 πŸ‘ 11 πŸ” 1 πŸ’¬ 0 πŸ“Œ 0
Post image

πŸ“£ [Openings] I'm now an Assistant Prof @westernu.ca CS dept. Funded PhD & MSc positions available! Topics: large probabilistic models, decision-making under uncertainty, and apps in AI4Science. More on agustinus.kristia.de/openings/

04.07.2025 22:55 πŸ‘ 18 πŸ” 5 πŸ’¬ 1 πŸ“Œ 1
Preview
Gaussian Processes and Reproducing Kernels: Connections and Equivalences This monograph studies the relations between two approaches using positive definite kernels: probabilistic methods using Gaussian processes, and non-probabilistic methods using reproducing kernel Hilb...

We've written a monograph on Gaussian processes and reproducing kernel methods (with @philipphennig.bsky.social, @sejdino.bsky.social and Bharath Sriperumbudur).

arxiv.org/abs/2506.17366

24.06.2025 08:35 πŸ‘ 36 πŸ” 11 πŸ’¬ 0 πŸ“Œ 0
ProbNum25 Coming

Registration for ProbNum 2025, to be held at EURECOM in southern France from 1st to 3rd September, is now open (deadline: 10th August). A tentative schedule and accepted papers are now online.

probnum25.github.io

04.06.2025 16:26 πŸ‘ 4 πŸ” 3 πŸ’¬ 0 πŸ“Œ 0
ProbNum25 Coming

Registration for ProbNum 2025, to be held at EURECOM in southern France from 1st to 3rd September, is now open (deadline: 10th August). A tentative schedule and accepted papers are now online.

probnum25.github.io

04.06.2025 16:26 πŸ‘ 4 πŸ” 3 πŸ’¬ 0 πŸ“Œ 0
Post image

Beautiful...

02.06.2025 07:18 πŸ‘ 3 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

Rio is looking forward to seeing all your works on #fast #reliable #tractable #models #circuits #tensor #networks #flows!

please share!

22.05.2025 05:26 πŸ‘ 6 πŸ” 8 πŸ’¬ 0 πŸ“Œ 0
Post image

Life goes on...

12.05.2025 08:25 πŸ‘ 2 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0
Post image

Bonjour...

09.05.2025 17:25 πŸ‘ 3 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0
Preview
A Dictionary of Closed-Form Kernel Mean Embeddings Kernel mean embeddings -- integrals of a kernel with respect to a probability distribution -- are essential in Bayesian quadrature, but also widely used in other computational tools for numerical inte...

Ever used kernel mean embeddings or MMD? Wished you had a closed-form expression so you wouldn't need to estimate these? See our new paper with Alex Gessner, Toni Karvonen and Maren Mahsereci, which comes with an associated Python package. arxiv.org/abs/2504.18830

29.04.2025 05:08 πŸ‘ 21 πŸ” 6 πŸ’¬ 1 πŸ“Œ 0
Post image

Why...?

24.04.2025 13:28 πŸ‘ 2 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0
Post image

"Donkeys and the Morning Sun" by Monet (2025)

24.04.2025 07:47 πŸ‘ 2 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

I didn't know that :) What's the countermeasure you invented?

24.04.2025 06:51 πŸ‘ 0 πŸ” 0 πŸ’¬ 1 πŸ“Œ 0
Post image

Why does the chain of sadness never end?

24.04.2025 05:36 πŸ‘ 7 πŸ” 0 πŸ’¬ 2 πŸ“Œ 0