computing Bayes, with some approximations

a journal of the conquest, war, famine, death, and chaos year

Had to find a fifth horseman of the Apocalypse for 2024, and figured out that Terry Pratchett was the ultimate expert on these characters, witness

“The Four Horsemen whose Ride presages the end of the world are known to be Death, War, Famine, and Pestilence. But even less significant events have their own Horsemen. For example, the Four Horsemen of the Common Cold are Sniffles, Chesty, Nostril, and Lack of Tissues; the Four Horsemen whose appearance foreshadows any public holiday are Storm, Gales, Sleet, and Contra-flow.” Terry Pratchet, Interesting Times

Hence followed suit when he signaled Chaos as the fifth (part-time) horseman. In Thief of Time. (Incidentally, I do not remember ever finishing a single book from his Discworld series, despite several attempts and encouragements from fellow readers.) With the precision of “not [being] the chaos which can be used to draw pretty Mandelbrot patterns”.

Read The Bone Season by Samantha Shannon. While I did not like much the author’s later The Priory of the Orange Tree, bought in one of the many Montréal bookstores and read by Lac Saint-Jean, which I found very shallow and predictable [the book, not the lake!], with most of the tropes of the genre (e.g., ninja-like fighters, heroes uncovering long-lost magical artefacts, super-evil entity about to return to life/power, a few predestined characters saving the Universe), unrealistic events, all-too-convenient coincidences, with little efforts put in the construction of the world, of the magical rules, or of the political structure, I still picked this rather heavy volume in a St Kilda bookstore, in [misguided] prevision for the flight back to Paris and read it later over the Winter solstice break. It is actually much better and enjoyable, in terms of world building for certain, involving a modified history of our World, mostly Britain and Ireland, with a non-human dominant species in the background, a no-man’s land gothic Oxford surrounded by zombies (?), and a part of humanity endowed with paranormal powers. The characters have some depth as well, even when remaining somehow predictable and often repetitive, and the scenario is mostly solid, apart from the unraveling finale. (The Guardian pointed out a direct filiation to Charlotte Brontë’s Jane Eyre, which I certainly did not see coming.) I am unsure I am ready for the next six volumes in the series, though!

Watched Gyeongseong, a Korean TV series that is a mix of horror, period, WW II, sadistic, mad scientists, nationalist, colonialist, family, survival, love, comical, soapy, teary stories all at once! Gyeongseong [경성, capital city] being the name of Seoul during the Japanese occupation and the entire series takes place in a recreated city that resembles the one in other period series I watched. Still enjoyable with a complete suspension of belief (as the scenario accumulates convenient coincidences, clueless Japanese occupiers, and bullet proof main characters) and the acceptance of the cartoonesque nature of fights and innumerable deaths.

futuristic statistical science [editorial]

This special issue of Statistical Science is devoted to the future of Bayesian computational statistics, from several perspectives. It involves a large group of researchers who contributed to collective articles, bringing their own perspectives and research interests into these surveys. Somewhat paradoxically, it starts with the past—and a conference on a Gold Coast beach. Martin, Frazier, and Robert first submitted a survey on the history of Bayesian computation, written after Gael Martin delivered a plenary lecture at Bayes on the Beach, a conference held in November 2017 in Surfers Paradise, Gold Coast, Queensland, and organised by Bayesian Research and Applications Group (BRAG), the Bayesian research group headed by Kerrie Mengersen at the Queensland University of Technology (QUT). Following a first round of reviews, this paper got split into two separate articles, Computing Bayes: From Then ‘Til Now , retracing some of the history of Bayesian computation, and Approximating Bayes in the 21st Century, which is both a survey and a prospective on the directions and trends of approximate Bayesian approaches (and not solely ABC). At this point, Sonia Petrone, editor of Statistical Science, suggested we had a special issue on the whole issue of trends of interest and promise for Bayesian computational statistics. Joining forces, after some delays and failures to convince others to engage, or to produce multilevel papers with distinct vignettes, we eventually put together an additional four papers, where lead authors gathered further authors to produce this diverse picture of some incoming advances in the field. We have deliberated avoided topics which have excellent recent reviews— such as Stein’s method, sequential Monte Carlo, piecewise deterministic Markov processes— and topics which are still in their infancy, such as the relationship of Bayesian approaches to large language models (LLMs) and foundation models.

Within this issue, Past, Present, and Future of Software for Bayesian Inference from Erik Štrumbelj & al covers the state of the art in the most popular Bayesian software, reminding us of the massive impact BUGS has had on the adoption of Bayesian tools since its early introduction in the early 1990s (which I remember discovering at the Fourth Valencia meeting on Bayesian statistics in April 1991). With an interesting distinction between first and second generations, and a light foray of the potential third generation, maybe missing the role of LLMs in coding that are already impacting the approach to computing and the less immediate revolution brought by quantum computing. Winter & al.’s The Future of Bayesian Computation [TITLE TO CHANCE] is making a link with machine learning techniques, without looking at the scariest issue of how Bayesian inference can survive in a machine learning world! While it produces an additional foray into the blurry division between proper sampling (à la MCMC) and approximations, additional to the historical Martin et al. (2024), it articulates these aspects within a (deep) machine learning perspective, emphasizing the role of summaries produced by generative models exploiting the power of neural network computation/optimization. And the pivotal reliance on variational Bayes, which is the most active common denominator with machine learning. With further entries on major issues like distributed computing, opening on the important aspect of data protection and guaranteed  privacy. We particularly like the clinical presentation of this paper with attention to automation and limitations. Normalizing flows actually link this paper with Heng, Bortoli and Doucet’s coverage of the Schrödinger bridge, which is a more focussed coverage of recent advances on possibly the next generation of posterior samplers. The final paper, Bayesian experimental design by Rainforth & al., provides a most convincing application of the methods exposed in the earlier papers in that the field of Bayesian design has hugely benefited from the occurrence of such tools to become a prevalent way of designing statistical experiments in real settings.

We feel the future of Bayesian computing is bright! The Monte Carlo revolution of the 1990s continues to be a huge influence on today’s work, and now is complemented by an exciting range of new directions informed by modern machine learning.

Dennis Prangle and Christian P Robert

a journal of the conquest, war, famine, and death [almost new] year

Read the fabulous Exhaltation over the trip to Australia [which deserved a post of its own!] and not much more, as I tried House of Earth and Blood. A terrible book I fortunately bought second-hand and which I returned to the seller at the South Melbourne market after one evening reading thru the first 100+ dreary pages. Terrible scenario, terrible style, with a caricature of characters behaving like college students without the least creativity in making a proper alternate universe for their acting. Hard to believe this could have been on the New York Times bestseller list….

Cooked more stir-fries both away and at home, returning to Fall veggies like cabbages and gourds, along local mushrooms and fortunately had to get rid of a large lot of kimchi, making kimchi mandu and kimchi fried rice for the first time. And miso glazed eggplant (nasu degaku) after eating some in Eke, the local Japanese restaurant. Sample fares at the great South Melbourne market, from fabulous Tasmanian oysters [albeit generating their own weight in plastic containers!] to a decent kouign-amann and a disappointing cannolo! Bringing home local fruits and fiery birdseye chili peppers.

Watched Vagabond, a conspiracy story set in modern South Korea, frankly  disappointing both in the lameness of the scenario, the obvious red herrings, the disregard for human life with endless executions, and the underlying racist tones for the scenes shot in North Africa and the Middle East…

St Kilda beach² [jatp]