finding our way in the dark

The paper Finding our Way in the Dark: Approximate MCMC for Approximate Bayesian Methods by Evgeny Levi and (my friend) Radu Craiu, recently got published in Bayesian Analysis. The central motivation for their work is that both ABC and synthetic likelihood are costly methods when the data is large and does not allow for smaller summaries. That is, when summaries S of smaller dimension cannot be directly simulated. The idea is to try to estimate

since this is the substitute for the likelihood used for ABC. (A related idea is to build an approximate and conditional [on θ] distribution on the distance, idea with which Doc. Stoehr and I played a wee bit without getting anything definitely interesting!) This is a one-dimensional object, hence non-parametric estimates could be considered… For instance using k-nearest neighbour methods (which were already linked with ABC by Gérard Biau and co-authors.) A random forest could also be used (?). Or neural nets. The method still requires a full simulation of new datasets, so I wonder at the gain unless the replacement of the naïve indicator with h(θ) brings clear improvement to the approximation. Hence much fewer simulations. The ESS reduction is definitely improved, esp. since the CPU cost is higher. Could this be associated with the recourse to independent proposals?

In a sence, Bayesian synthetic likelihood does not convey the same appeal, since is a bit more of a tough cookie: approximating the mean and variance is multidimensional. (BSL is always more expensive!)

As a side remark, the authors use two chains in parallel to simplify convergence proofs, as we did a while ago with AMIS!

a journal of the plague year² [600+]

Returned to Warwick for the first time in 600 days!, most pleasantly reconnecting with my colleagues there and realising some were almost as freshly back to the department as myself. But also noticed a strong difference with France in terms of wearing mask and practicing social distanciation among the students (at the University) and the general population (in the local Tesco or the train to the airport). Which may explain for the persistently high number of contaminations, when compared with neighbouring countries. Despite its high vaccination rate.

Read the second volume of Baru Cormorant, after enjoying so much the first instalment (under the Corsican relentless sun). However, it was such a disappointment, as it seemed written by a completely different author, including the style, with the story being more broken, more difficult to follow, and the characters becoming shallow and uninteresting. This is particularly true of Baru, who sent from a sharp focus on her goal in Traitor, to a purposeless floatsam in Monster. With a highly artificial rescue of her ethics in [warning, spoiler ahead!] killing her lover Tain Hu (and hundreds of others) in the first volume. I currently doubt I will buy the third one… A stormy night kept me awake and as a result help me finish the last hundred pages!

Watched Mute, dubbed a “Netflix disaster” in The Guardian! (and following the appalling Warcraft!). Which postulate of a mute character could have been rewarding, had not the characters be of cardboard consistency. And the plot so transparent most of the scenes had to be shot at night. And the cheap plagiarism of Blade Runner is simply ludicrous. I also watched Bong Joon-Ho’s Okja, a satire about the meat industry producing hippo-like pigs that I found very heavy-handed, especially in its characters.Started cooking bread on a regular basis once again, as weather is turning colder and baking also warms up our drafty kitchen! Now turning to a heavier type of loaf, mixed wheat and rye, and reducing the amount of water to make it last longer. And also made a first attempt at Okonomyaki, this tortilla-like dish made of shredded cabbage and flour that I enjoyed very much in Osaka. The result was pleasant enough but very, very far from the original, maybe due to my using (soba) buckwheat rather than plain flour. Or using regular stock rather than daishi stock.

stranded

Bangalore snapshot [ಬೆಂಗಳೂರು ಚಿತ್ರ]

 

Bangalore snapshot [ಬೆಂಗಳೂರು ಚಿತ್ರ]