another fine…

bye, sempé…

`Paris is in anarchy’ [cycle woes]

An overblown view of the cycling war in Paris, from New York! I read with amusement the report on how Xing a Parisian street is a matter of life or death, when anarclists go through red lights while shouting at pedestrians… Actually, the figures show that the number of accidents involving cyclists (as victims or culprits) has only gone up by 30% when the traffic has increased by 70%. And I could not find an online trace of a pedestrian killed by a cyclist over the past years. Based on my weekly 130 kilometer biking average, mostly to and from Paris Dauphine, I do not perceive a major tension between pedestrians and cyclists, maybe because I am not entering the centre of town (and give priority to pedestrians at both green and red lights). The danger in my experience comes rather from other cyclists’ unpredictable paths, (psychopath) mopeds that run on cycle paths, and cars turning right without checking for bicycles. But I concur with the point made in this article of a poor network of cycle paths, with too many discontinuities, bad surface, inexistent maintenance (esp. in winter months when wet leaves accumulate there and all year long for broken glass and metal parts), and the deadly pavés! Which are unpleasant for road bikes (ask the Paris-Roubaix runners!), slippery, esp. when frosted (speaking from experience), and damaging to tubes and ties. As it happens, I have had thee tube punctures over the three past weeks, two of which were due to running over a particularly uneven pavé or entering a cycle path with a very high step. (And a total of six since April. Making me reconsider using an heavier mountain bike instead. After switching unsuccessfully to anti-puncture road tyres…)

cycle static

Heard while (safely and slowly) passing a group of pedestrians on a shared green-way, south of Paris:

“Les gens un jour je vais tous les tuer.”

which roughly translates as “One of these days, I will kill all people.” And whose meaning (if any) escapes me.

EM degeneracy

At the MHC 2021 conference today (to which I biked to attend for real!, first time since BayesComp!) I listened to Christophe Biernacki exposing the dangers of EM applied to mixtures in the presence of missing data, namely that the algorithm has a rising probability to reach a degenerate solution, namely a single observation component. Rising in the proportion of missing data. This is not hugely surprising as there is a real (global) mode at this solution. If one observation components are prohibited, they should not be accepted in the EM update. Just as in Bayesian analyses with improper priors, the likelihood should bar single or double  observations components… Which of course makes EM harder to implement. Or not?! MCEM, SEM and Gibbs are obviously straightforward to modify in this case.

Judith Rousseau also gave a fascinating talk on the properties of non-parametric mixtures, from a surprisingly light set of conditions for identifiability to posterior consistency . With an interesting use of several priors simultaneously that is a particular case of the cut models. Namely a correct joint distribution that cannot be a posterior, although this does not impact simulation issues. And a nice trick turning a hidden Markov chain into a fully finite hidden Markov chain as it is sufficient to recover a Bernstein von Mises asymptotic. If inefficient. Sylvain LeCorff presented a pseudo-marginal sequential sampler for smoothing, when the transition densities are replaced by unbiased estimators. With connection with approximate Bayesian computation smoothing. This proves harder than I first imagined because of the backward-sampling operations…