Archive for BayesComp 2025

« Older Entries

ISBA⁴

Posted in Books, pictures, Running, Statistics, Travel, University life with tags , , , , , , , , , , , , , , , , on July 3, 2026 by xi'an

After a 5am run in emptied and very humid Nagano, to tour the castle grounds, if not the castle itself, with a rather unpleasant urban route, and a recovery matcha scone at the local French (-inspired) bakery, resisting the kouign aman!, I attended the highly diverse Bayesian non-parametric session (with mentions of Lisieux, near my childhood grounds, if in relation with English history, since the bishop of Lisieux appeared in Geoff Nicholls’ graph analysis. And Gregor Katzner’s repulsive mixtures.)

Then skipped Peter Muller’s de Finetti talk [sorry Peter!] with a few friends to get to nearby Atsuta-jingū (Shinto shrine) and the nearby Bunka-den, with a famous hitsumabushi (eel) restaurant that was alas fully booked. But a slightly less fancy one delivered well enough. This was a nice place, with a large park, full of worshippers clapping and touring the shrines. We even X’ed a group of salarymen in their company uniform welcomed by a Shinto priest.

And I was back in time for the junior Bayesian (jISBA) session. With Deborah Sulem talking on statistical fairness, which I could not fully distinguish from the general issue of missing covariates (corresponding to discriminated categories) or from unbalanced sampling. Jack Jewson (also coauthor of the paper with Deborah) argued about using quasi-likelihoods for consistent model selection, presumably most convincingly but jetlag hit during his talk and I had a fit of micro-naps… And Noirrit Chandra discussed a privacy-preserving Bayesian nonparametric approach through clusters and the release of MCMC outputs, close enough to my own research to fight the naps. The day closed with Emtiyaz Khan talking as at BayesComp last year on acceleration techniques as sustainable AI (although I am afraid this does not tackle the unsustainability of AI as a whole).

Leave a comment »

approximately Bayes [on Skye]

Posted in Mountains, pictures, Running, Statistics, Travel, University life with tags , , , , , , , , , , , , , , , , , , , , , , , , , , , , on May 28, 2026 by xi'an

Wow, what an exciting workshop in an equally exciting place! Strong themes were post-Bayes (Gibbs priors, martingale priors, predictive Bayes, &tc.) and deep neural network modelling. With animated discussions allowed by the free windows planned in the program. And the very early dinner at Sabhal Mòr Ostaig that let a long sunlit evening for impromptu Q&A’s [with a serving of lamb and another of haggis pie!]. Making me realise the large corpus of work I had missed in the past years on these topics, even though the satellite of BayesComp last year was already an eye opener. (Stay tuned for news about BayesComp 2027 & its mirror in Aussois!) The proposal in Jeff Miller’s discussion of Jeremias Knoblauch’s overview of post-Bayes [I’d rather favour another name!]  to consider directly likelihood values as the data was particularly appealing to me, while reminding me of the foundations of nested sampling. (Hopefully, a new perspective on uncertainty assessment for nested sampling is soon to be completed!)

On the non-academic side, the long days in The North helped with my running with above 90km bagged in the week (and no downpour on the runs). But little to my swimming since the water was cold enough to limit my laps to 5mn each time! Paradoxically the worst day was the one I chose for climbing the Inaccessible Pinnacle (as expanded in another ‘Og entry).

Leave a comment »

Nature on mirror conferences

Posted in Books, Statistics, University life with tags , , , , , , , , , , , , , , , , , , , , , , , on January 28, 2026 by xi'an

In a news article entitled “Scientists skip key US meetings — and seize on smaller alternatives”, Nature discusses the impact of the restrictive policies put in place by the Trump administration on US conferences and their attendance. Including the multiplication of mirror and satellite meetings. One of the examples in the article is Neurips 2025,

“…the artificial-intelligence conference NeurIPS hosted not only its main meeting in San Diego, California, but also its first-ever alternative location, in Mexico City, with the goal of alleviating travel challenges (…) in response to “skyrocketing attendance and difficulties in obtaining travel visas some attendees have experienced in the past few years when only one location was available” [while] a group of AI researchers in Europe organized an independent spin-off conference, dubbed EurIPS, in Copenhagen (…) owing to concerns including climate change [and people expressing] a desire for a less hostile environment”

With a limited number of 500 participants attending in Mexico. And a massive number in Copenhagen, over 2,000! With a final quote from Emtiyaz Khan (a plenary speaker at ISBA 2026):

“[I] chose to travel to EurIPS rather than NeurIPS because of the difficulties many others faced in getting into the United States. The smaller nature of EurIPS made it much easier to meet and interact with other scientists. I absolutely loved it and I would love to see it happen again.”

This state of affairs is not going to vanish with Trump adding more countries to the banned country list, 75 at this stage!, and this is a call to arms for ISBA and IMS conference organisers towards planning for multi-hub configurations, since such international organisations cannot exclude a third of the countries in the World from attending their conferences. Which makes our current ISBA survey all the more relevant! I am currently building a mirror meeting for BayesComp 2027 in Aussois, French Alps. For those who cannot or do not wish to travel to Texas for the main conference.

Leave a comment »

BayesComp 2025.4

Posted in pictures, Running, Statistics, Travel, University life with tags , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , on June 21, 2025 by xi'an

The third and final day of the (main) conference started tih Emtiyaz Khan’s plenary talk on adaptive Bayesian intelligence. Or, imho, [adaptive [Bayesian]] intelligence, with the brackets indicating redundancy since intelligence need include adaptivity and [intelligent] adaptivity need proceed in a Bayesian way! Focussing first on the Bayesian learning rule via variational Bayes (with a stress on Kingma’s 1994 Adam optimisation algorithm, the “most cited paper” [in machine learning]) where learning boils down to gradient steps (due to the exponential family structure), themselves versions of Taylor (or Laplace) approximations). With an interesting vision of Bayesian updating as accounting for prediction mismatch. (I missed the connection Roberta in IMDb appearing in one slide!)

 The following session offered no dilemma [sorry, Alex, Axel, Chris, Robert, Sumeet, Victor!] since it included the federated learning session I organised, with Louis Asslet, Conor Hassan, and Jean-Michel Marin as speakers. Louis’ talk was on confidential [homomorphic] accept-reject algorithms to learn from other sources, while preserving (differential?) privacy, part of which came during Les Houches workshops I organised this Spring and the one before. Exploiting the additive features of log-likelihoods and exponential variates and adopting a testing perspective on privacy. Conor motivated his model with the Australian cancer atlas project Kerrie Mengersen and others have been developing over the years. The federated approach relies on variational approximations that return the same answer as an exact resolution, but more efficiently. (From a privacy perspective, I wonder at the impact of variational approximations on protecting the data, which boils down to a choice of (sufficient) statistics for the exponential families behind those approximations.) For more complicated models incorporating spatial dependence prohibits full Bayesian inference, unfortunately. Jean-Michel commented on the richness of methods for simulation-based inference, incl. model choice. His focus was on using sequential neural likelihood estimation and sequential importance sampling to approximate evidence. As in the Read Paper of Del Moral et al. (2006). Mentioning a neural version of the harmonic mean estimator by Spurio Mancini et al.  (2023)! I wondered at the degree of (Rao-Blackwell) recycling involved in the computation, Jean-Michel’s answer being that AMIS is soon coming [in a theatre near you!].

The afternoon sessions did offer any reprieve in the choice of topic! I first went to Approximate Methods for Accelerated Sampling, with Rong Tang evaluating the informativeness of summary statistics through a divergence evaluation. Using autoencoders to replace the intractable posterior, with sliced minimal model discrepancy (MMD) and (pseudo?) score matching loss for divergences (reminding me of indirect inference and synthetic likelihood). Yun Yang discussed a variational proposal to estimate the number of components in a mixture model. Surprising given the multimodal structure of mixture posteriors. And the overall irregularity of (evil!) mixture models. But I could not figure out from the talk the form of the approximation.

On the food scene, tasted a nice and spicy Peranakan rice vermicelli dish called Mee Siam yesterday in a campus restaurant, which sustained me fore the rest of the day, including the ABC s/webinar. And another spicy hot pot today at NUS, to catch up on veggies, while missing the chili crab local specialty on that trip.

Leave a comment »

BayesComp 2025.3

Posted in pictures, Running, Statistics, Travel, University life with tags , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , on June 20, 2025 by xi'an

The second day of the conference started with a cooler and less humid weather (although this did not last!), although my brain felt a wee bit foggy from a lack of sleep (and I almost crashed while running on the hotel treadmill, at 14.5km/h!), and the plenary talk of my friend of many years Sylvia Früwirth-Schnatter on horseshoe priors and time-varying time series (à la West). With a nice closed-form representation involving hypergeometric functions of the second kind (my favourite!), with the addition of a triple-Gamma prior. Sylvia stressed on the enormous impact of the prior choice on change-point detection, which was already the point in the original horseshoe paper (as opposed to George’s Lasso prior). Without incorporating any specific modelling on potential change-point, fair enough given that the parameter is moving with time, unhindered. Her MCMC choices involved discrete parameters with Negative Binomial and Poisson parameters, allowing for partially integrated or collapsed solutions. Possibly further improved by Swendsen-Wang steps.

I then attended the (advanced) Langevin session after agonising upon my choice for a wealth of options! Sam Power presented a talk linking simulation with optimisation targets, over measure spaces. With Wasserstein gradient flow algorithms that resemble Langevin algorithms once discretised by a particle system. (A natural resolution producing a somewhat unnatural form of measure estimator since made of Dirac masses, from which very little can be learned.) Then [my Warwick colleague & coauthor] Any Wang on underdamped Langevin diffusions. when Poincaré‘s inequality fails, but convergence (in total variation) still occurs. Followed by Peter Whalley on splitting methods (where random hypergeometric subsampling dominates Robbins-Monro) and stochastic gradient algorithms, in a connected (to the previous talks) way since involving underdamped aspects. (With a personal discovery of Polyak’s heavy ball method.)

The afternoon session saw me facing a terrible dilemma with three close friends talking at the same time! Eventually opting for PDMPs, over simulation-based inference and recalibration for approximate Bayesian methods. Kengo Kamatani gave a general introduction to PDMPs, before explaining the automated implementation he considered with Charly Andral (during Charly’s visit to ISM, Tokyo, two summers ago). Towards accelerating the generation of the jump time. Then Luke Hardcastle applied PDMPs for survival prediction, using spike & slab priors and sticky PDMPs. And Jere Koskela (formerly Warwick) extended zig-zag sampling to discrete settings (incl. Kingman’s coalescent.)

The (rather long) day was not over yet since we had planned an extra on-site OWABI seminar & webinar with two participants in the conference, Filippo Pagani (Warwick and OCEAN postdoc) using fusion for federated learning, with a trapezoidal approximation, and Maurizio Filippone on GANs as hidden perfect ABC model selection, a GAN providing an automatic density estimator… With astounding Gemini-generated cartoons! Videos are soon to be available. A big congrats to the speakers who managed to convey their ideas and results despite the late hour! (On the extra-academic side, I was invited last night to a genuine Szechuan dinner in Chinatown, with a large array of spicy dishes if not that spicy!, and a rare opportunity to taste abalone. And bullfrogs. Quite a treat! And a good reason to skip dinner altogether!)

Leave a comment »

« Older Entries