Archive for Elsevier

Advancements in Bayesian Methods and Implementations [to appear]

Posted in Books, Statistics with tags , , , , , on July 17, 2022 by xi'an

As noted in another post, I wrote a chapter on Bayesian testing for an incoming handbook, Advancements in Bayesian methods and implementations which is published by Elsevier at an atrocious price (as usual). Here is the table of contents:

1. Fisher Information, Cramèr-Rao and Bayesian Paradigm by Roy Frieden
2. Compound beta binomial distribution functions by Angelo Plastino
3. MCMC for GLMMS by Vivekananda Roy
4. Signal Processing and Bayesian by Chandra Murthy
5. Mathematical theory of Bayesian statistics where all models are wrong by Sumio Watanabe
6. Machine Learning and Bayesian by Jun Zhu
7. Non-parametric Bayes by Stephen Walker
8. [50 shades of] Bayesian testing [of hypotheses] by Christian P. Robert
9. Data Analysis with humans by Sumio Kaski
10. Bayesian Inference under selection by G. Alastair Young
10. Variational inference or Functional horseshoe by Anirban Bhattacharya
11. Generalized Bayes by Ryan Martin

and my chapter is also available on arXiv, quickly gathered from earlier short courses at O’Bayes meetings and some xianblog entries on the topic, hence not containing much novelty!

reXing the bridge

Posted in Books, pictures, Statistics with tags , , , , , , , , , on April 27, 2021 by xi'an

As I was re-reading Xiao-Li  Meng’s and Wing Hung Wong’s 1996 bridge sampling paper in Statistica Sinica, I realised they were making the link with Geyer’s (1994) mythical tech report, in the sense that the iterative construction of α functions “converges to the `reverse logistic regression’  described in Geyer (1994) for the two-density cases” (p.839). Although they also saw the later as an “iterative” application of Torrie and Valleau’s (1977) “umbrella sampling” estimator. And cited Bennett (1976) in the Journal of Computational Physics [for which Elsevier still asks for $39.95!] as the originator of the formula [check (6)]. And of the optimal solution (check (8)). Bennett (1976) also mentions that the method fares poorly when the targets do not overlap:

“When the two ensembles neither overlap nor satisfy the above smoothness condition, an accurate estimate of the free energy cannot be made without gathering additional MC data from one or more intermediate ensembles”

in which case this sequence of intermediate targets could be constructed and, who knows?!, optimised. (This may be the chain solution discussed in the conclusion of the paper.) Another optimisation not considered in enough detail is the allocation of the computing time to the two densities, maybe using a bandit strategy to avoid estimating the variance of the importance weights first.

Nature tea[dbits]

Posted in Books, pictures, University life, Wines with tags , , , , , , , , , , , , , , , , on February 28, 2019 by xi'an

A very special issue of Nature (7 February 2019, vol. 556, no. 7742). With an outlook section on tea, plus a few research papers (and ads) on my principal beverage. News about the REF, Elsevier’s and Huawei’s woes with the University of California, the dangerous weakening of Title IX by the Trump administration, and a long report on the statistical analysis of Hurricane Maria deaths, involving mostly epidemiologists, but also Patrick Ball who took part in our Bayes for Good workshop at CIRM. Plus China’s food crisis and ways to reduce cropland losses and food waste. Concerning the tea part(y), a philogenetic study of different samples led to the theory that tea was domesticated thrice, twice in Yunnan (China) and once in Assam (India), with a divergence estimated at more than twenty thousand years ago. Another article on Pu-Ehr, with the potential impacts of climate change on this very unique tea. With a further remark that higher altitudes increase the anti-oxydant level of tea… And a fascinating description of agro-forestry where tea and vegetables are grown in a forest that regulates sun exposure, moisture evaporation, and soil nutrients.

preprints promote confusion and distorsion, and don’t blame journalists!

Posted in Books, pictures, Travel, University life with tags , , , , , , , , on October 4, 2018 by xi'an

“…anyone considering publicizing a preprint have a responsibility.”

On my way to the airport, flying to B’ham, I read an older issue of Nature that contained this incredible editorial entry from Tom Sheldon Tim Horton, calling for regulation of preprints or worse, for the reason that journalists could misunderstand their contents and over-hype a minor or worse wrong claim. Taking as mistaken illustration the case of the Séralini et al. paper, about the Monsanto maize, which happened to be published under “embargo” conditions and reproduced in most media before a scientific storm erupted on the lack of significance of the samples. This call is unbelievably cheeky and downright absurd as it shifts the responsibility away from the journalists to the scientific community, throwing the “check your sources” principle of investigative journalism down the drain. As if the only reason for immediately publishing front-page discoveries is not to beat the competition and attract more readers…

The irony of seeing this piece in Nature is that a few pages later, there is a news entry on German and Swedish institutions breaking negotiations with Elsevier, as the publisher refuses to join a global package of open source publications. Nothing seems amiss about this nice aspect of scientific publishing with the author of this editorial, nor with the further reports of retraction of published paper in the same issue. Presumably because journalists have already moved to the next hot discovery by the time the retractions at last appear…! And to answer the final question of “Should all preprints be emblazoned with a warning aimed at journalists that work has not been peer reviewed?”, no, no, and no: preprints are not written for journalists or the general public. Unsurprisingly, the tribune induced outraged reactions from Nature readers.

Nature snapshot

Posted in Statistics with tags , , , , , , , on March 5, 2017 by xi'an

The recent issue of Nature, as of Jan 26, 2017!, contained a cartload of interesting review and coverage articles, from the latest version of the quantum computer D-Wave, with a paragraph on quantum annealing that reminded me of a recent arXiv paper I could not understand, seemingly turning the mathematical problem of multivariate optimisation into a truly physical process, to the continuing (Nature-wise) debate on how to oppose Trump, to the biases and shortcomings of policing software, with a mention of Lum and Isaac I discussed here a few months ago, to the unsuspected difficulty to publish a referee’s report when the publisher is Elsevier (unsuspected and unsurprising!)—although I know of colleagues and authors disapproving my publishing referee’s reports identified as such—, to an amazing picture of a bundle of neurons monitored simultaneously, to an entry in the career section on scientific computing and the importance of coding for young investigators, with R at the forefront!

%d bloggers like this: