Arthur Gretton

Looking forward to next week's Winter School on Causality and Explainable AI!

xai-winter-school.github.io

Hope to see you there!

Fantastic news, congratulations!!

Thank you for a fantastic conference!

Sequential kernel embedding for mediated and time-varying dose response curves

Appearing in Bernoulli:
projecteuclid.org/journals/ber...

with preprint here: arxiv.org/abs/2111.03950

...along with code!
github.com/liyuan9988/K...

Rahul Singh, Liyuan Xu

UCL – University College London UCL is consistently ranked as one of the top ten universities in the world (QS World University Rankings 2010-2022) and is No.2 in the UK for research power (Research Excellence Framework 2021).

Research Fellow position open at @gatsbyucl.bsky.social
to work with me and Jason Hartford on Causality in Biological Systems!

Apply at link, deadline is 27 August:
www.ucl.ac.uk/work-at-ucl/...

The method accepts draft proposals sequentially - once a proposal is rejected, a maximal coupling is used to obtain a valid sample, and the process repeats.

re "still working with kernels" - see the other ICML 2025 paper, arxiv.org/abs/2502.02483 which uses distributional kernel scoring rules!

Accelerated Diffusion Models via Speculative Sampling, at #icml25 !

16:30 Tuesday July 15 poster E-3012

arxiv.org/abs/2501.05370

@vdebortoli.bsky.social Galashov @arnauddoucet.bsky.social

Distributional diffusion models with scoring rules at #icml25

Fewer, larger denoising steps using distributional losses!

Wednesday 11am poster E-1910

arxiv.org/pdf/2502.02483

@vdebortoli.bsky.social
Galashov Guntupalli Zhou
@sirbayes.bsky.social
@arnauddoucet.bsky.social

Distributional Reduction paper with H. Van Assel, @ncourty.bsky.social, T. Vayer , C. Vincent-Cuaz, and @pfrossard.bsky.social is accepted at TMLR. We show that both dimensionality reduction and clustering can be seen as minimizing an optimal transport loss 🧵1/5. openreview.net/forum?id=cll...

Composite Goodness-of-fit Tests with Kernels, now out in JMLR!

www.jmlr.org/papers/v26/2...

Test if your distribution comes from ✨any✨ member of a parametric family. Comes in MMD and KSD flavours, and with code.

@oscarkey.bsky.social @fxbriol.bsky.social Tamara Fernandez

Turns out that overfitting is the right approach when you want to generalize to new tasks!

Mattes Mollenhauer, Nicole M\"ucke, Dimitri Meunier, Arthur Gretton: Regularized least squares learning with heavy-tailed noise is minimax optimal https://arxiv.org/abs/2505.14214 https://arxiv.org/pdf/2505.14214 https://arxiv.org/html/2505.14214

Looking forward to this!

Kernel Single Proxy Control for Deterministic Confounding

at #AISTATS25

Proxy causal learning generally requires two proxy variables - a treatment and an outcome proxy. When is it possible to use just one?

arxiv.org/abs/2308.04585

Liyuan Xu

Credal Two-Sample Tests of Epistemic Uncertainty
at #AISTATS25

Compare credal sets: convex sets of prob measures where elements capture aleatoric uncertainty; set represents epistemic uncertainty.

arxiv.org/abs/2410.12921

@slchau.bsky.social Schrab @sejdino.bsky.social @krikamol.bsky.social

Spectral Representation for Causal Estimation with Hidden Confounders
at #AISTATS2025

A spectral method for causal effect estimation with hidden confounders, for instrumental variable and proxy causal learning
arxiv.org/abs/2407.10448

Haotian Sun, @antoine-mln.bsky.social, Tongzheng Ren, Bo Dai

Density Ratio-based Proxy Causal Learning Without Density Ratios 🤔

at #AISTATS2025

An alternative bridge function for proxy causal learning with hidden confounders.
arxiv.org/abs/2503.08371
Bozkurt, Deaner, @dimitrimeunier.bsky.social, Xu

Announcing : The 2nd International Summer School on Mathematical Aspects of Data Science
mathsdata2025.github.io
EPFL, Sept 1–5, 2025

Speakers:
Bach @bachfrancis.bsky.social
Bandeira
Mallat
Montanari
Peyré @gabrielpeyre.bsky.social

For PhD students & early-career researchers
Apply before May 15!

Optimality and Adaptivity of Deep Neural Features for Instrumental Variable Regression
#ICLR25

openreview.net/forum?id=ReI...

NNs
✨better than fixed-feature (kernel, sieve) when target has low spatial homogeneity,
✨more sample-efficient wrt Stage 1

Kim, @dimitrimeunier.bsky.social, Suzuki, Li

Deep MMD Gradient Flow Without Adversarial Training
at #ICLR2025

openreview.net/forum?id=Pf8...

Do you have a GAN critic? Then you have a diffusion!

Adaptive MMD gradient flow trained on a forward diffusion, competitive performance on image generation!

Galashov, @vdebortoli.bsky.social

Looking forward to this!

congratulations!!

Variance-Aware Estimation of Kernel Mean Embedding An important feature of kernel mean embeddings (KME) is that the rate of convergence of the empirical KME to the true distribution KME can be bounded independently of the dimension of the space, prope...

Our joint paper with Geoffrey Wolfer @gwolfer.bsky.social "Variance-Aware Estimation of the Kernel Mean Embedding" accepted for publication in the Journal of Machine Learning Research 🥳

arxiv.org/abs/2210.06672

Congratulations @lestermackey.bsky.social !!

Hey ELLIS PhD students, need to travel but low on funds? Learn how ELSA can help with that: bit.ly/4kqjyel

#ELLISPhD #MobilityFund #SustainableAI #ProjectsBuildingOnELLIS

I already advertised for this document when I posted it on arXiv, and later when it was published.

This week, with the agreement of the publisher, I uploaded the published version on arXiv.

Less typos, more references and additional sections including PAC-Bayes Bernstein.

arxiv.org/abs/2110.11216

The slides of my talk at OIST are online: pierrealquier.github.io/slides/okina...

Thanks to the organisers, and thanks to Frank Nielsen for the photo of my talk 🙏

Link to the paper: arxiv.org/abs/2412.18539

Video now public from the talk

"Learning to act in noisy contexts using deep proxy learning"

at the NeurIPS'24 Workshop on Causal Representation Learning!

Video:
neurips.cc/virtual/2024...

Slides:
www.gatsby.ucl.ac.uk/~gretton/cou...

super happy about this preprint! we can *finally* perform efficient exploration and find near-optimal stationary policies in infinite-horizon linear MDPs, and even use it for imitation learning :) working with @neu-rips.bsky.social and @lviano.bsky.social on this was so much fun!!