Archive for PhD students

BAYSM 2020, Kunming, China [reposted]

Posted in Kids, Mountains, pictures, Statistics, Travel, University life with tags , , , , , , , , on January 13, 2020 by xi'an

The 5th Bayesian Young Statisticians Meeting, BAYSM2020, will take place in Kunming, China (June 26-27, 2020) as a satellite to the ISBA 2020 world meeting. BAYSM is the official conference of j-ISBA, the junior section of the International Society for Bayesian Analysis. It is intended for Ph.D. Students, M.S. Students, Post-Docs, Young and Junior researchers working in the field of Bayesian statistics, providing an opportunity to connect with the Bayesian community at large. Senior discussants will be present at each session, providing participants with hints, suggestions and comments to their work. Distinguished professors of the Bayesian community will also participate as keynote speakers, making an altogether exciting program.

Registration is now open (https://baysm2020.uconn.edu/registration) and will be available with an early bird discount until May 1, 2020. The event will be hosted at the Science Hall of Yunnan University (Kunming, China) right before ISBA 2020 world meeting. BAYSM 2020 will include social events, providing the opportunity to get to know other junior Bayesians.

Young researchers interested in giving a talk or presenting a poster are invited to submit an extended abstract by March 29, 2020. All the instructions for the abstract submission are reported at the page https://baysm2020.uconn.edu/call-dates

Thanks to the generous support of ISBA, a number of travel awards are available to support young researchers.

Keynote speakers:
Maria De Iorio
David Dunson
Sylvia Frühwirth-Schnatter
Xuanlong Nguyen
Amy Shi
Jessica Utts

Confirmed discussants:
Jingheng Cai
Li Ma
Fernando Quintana
Francesco Stingo
Anmin Tang
Yemao Xia

While the meeting is organized for and by junior Bayesians, attendance is open to anyone who may be interested. For more information, please visit the conference website: https://baysm2020.uconn.edu/

postgraduate open day at Warwick [4 Dec]

Posted in pictures, Statistics, University life with tags , , , , , , , , , , on November 12, 2019 by xi'an

The department of Statistics at the University of Warwick is holding an open day for prospective PhD students on 4 December 2019, starting at 2pm (with free lunch at 1pm). In the Mathematical Sciences Building common room (room MB1.02). The Director of Graduate Studies, Professor Mark Steel, and the PhD admissions tutors Professors Martyn Plummer and Barbel Finkelstadt Rand will give short presentations about what it means to do a PhD, what it means to do it at Warwick, the benefits of a PhD degree, and the application process.

Subsequently there will be an informal meeting, during which students have the possibility to ask questions and find out more about the different PhD opportunities at Warwick Statistics; in fact, we offer a very broad range of possibilities, giving a lot of choice for potential applicants. Current members of staff will be invited to participate, to discuss potential projects.

UK travel expenses will be covered by the Department of Statistics (standard class travel by public transport with pre-booked tickets). Please register if interested in this event.

y a plus de mouchoirs au bureau des pleurs

Posted in pictures, University life with tags , , , , , , , , , on January 10, 2019 by xi'an

Once the French government started giving up to some requests of the unstructured “gilets jaunes” protesters, it was like a flood or flush gate had opened and every category was soon asking for a rise (in benefits) and a decrease (in taxes) or the abolition of a recent measure (like the new procedure for entering university after high school). As an illustration, I read a rather bemusing tribune in Le Monde from a collective of PhD students against asking non-EU students (including PhD students) to pay fees to study in French universities. This may sound a bit of a surrealistic debate from abroad, but the most curious point in the tribune [besides the seemingly paradoxical title of students against Bienvenue En France!] is to argue that asking these students to pay the intended amount would bring their net stipends below the legal minimum wage, considering that they are regular workers… (Which is not completely untrue when remembering that in France the stipends are taxed for income tax and retirement benefits and unemployment benefits, meaning that a new PhD graduate with no position can apply for these benefits.) It seems to me that the solution adopted in most countries, namely that the registration fees are incorporated within the PhD grants, could apply here as well… The other argument that raising these fees from essentially zero to 3000 euros is going to stop bright foreign students to do their PhD in France is not particularly strong when considering that the proportion of foreign students among PhD students here is slightly inferior to the proportion in the UK and the US, where the fees are anything but negligible, especially for foreign students.

coordinate sampler as a non-reversible Gibbs-like MCMC sampler

Posted in Books, Kids, Statistics, University life with tags , , , , , , , , on September 12, 2018 by xi'an

In connection with the talk I gave last July in Rennes for MCqMC 2018, I posted yesterday a preprint on arXiv of the work that my [soon to defend!] Dauphine PhD student Changye Wu and I did on an alternative PDMP. In this novel avatar of the zig-zag sampler,  a  non-reversible, continuous-time MCMC sampler, that we called the Coordinate Sampler, based on a piecewise deterministic Markov process. In addition to establishing the theoretical validity of this new sampling algorithm, we show in the same line as Deligiannidis et al.  (2018) that the Markov chain it induces exhibits geometrical ergodicity for distributions which tails decay at least as fast as an exponential distribution and at most as fast as a Gaussian distribution. A few numerical examples (a 2D banana shaped distribution à la Haario et al., 1999, strongly correlated high-dimensional normals, a log-Gaussian Cox process) highlight that our coordinate sampler is more efficient than the zig-zag sampler, in terms of effective sample size.Actually, we had sent this paper before the summer as a NIPS [2018] submission, but it did not make it through [the 4900 submissions this year and] the final review process, being eventually rated above the acceptance bar but not that above!

Hamiltonian tails

Posted in Books, Kids, R, Statistics, University life with tags , , , , , , on July 17, 2018 by xi'an

“We demonstrate HMC’s sensitivity to these parameters by sampling from a bivariate Gaussian with correlation coefficient 0.99. We consider three settings (ε,L) = {(0.16; 40); (0.16; 50); (0.15; 50)}” Ziyu Wang, Shakir Mohamed, and Nando De Freitas. 2013

In an experiment with my PhD student Changye Wu (who wrote all R codes used below), we looked back at a strange feature in an 2013 ICML paper by Wang, Mohamed, and De Freitas. Namely, a rather poor performance of an Hamiltonian Monte Carlo (leapfrog) algorithm on a two-dimensional strongly correlated Gaussian target, for very specific values of the parameters (ε,L) of the algorithm.

The Gaussian target associated with this sample stands right in the middle of the two clouds, as identified by Wang et al. And the leapfrog integration path for (ε,L)=(0.15,50)

keeps jumping between the two ridges (or tails) , with no stop in the middle. Changing ever so slightly (ε,L) to (ε,L)=(0.16,40) does not modify the path very much

but the HMC output is quite different since the cloud then sits right on top of the target

with no clear explanation except for a sort of periodicity in the leapfrog sequence associated with the velocity generated at the start of the code. Looking at the Hamiltonian values for (ε,L)=(0.15,50)

and for (ε,L)=(0.16,40)

does not help, except to point at a sequence located far in the tails of this Hamiltonian, surprisingly varying when supposed to be constant. At first, we thought the large value of ε was to blame but much smaller values still return poor convergence performances. As below for (ε,L)=(0.01,450)

computer strategies for complex Bayesian models

Posted in Books, Kids, Statistics, University life with tags , , , , , , , , on July 18, 2016 by xi'an

frontThis is the cover page of Marco Banterle‘s thesis, who will defend on Thursday [July 21, 13:00], at a rather quiet time for French universities, which is one reason for advertising it here. The thesis is built around several of Marco’s papers, like delayed acceptance, dimension expansion, and Gaussian copula for graphical models. The defence is open to everyone, so feel free to join if near Paris-Dauphine!

the penalty method

Posted in Statistics, University life with tags , , , , , , , , , , on July 7, 2016 by xi'an

“In this paper we will make conceptually simple generalization of Metropolis algorithm, by adjusting the acceptance ratio formula so that the transition probabilities are unaffected by the fluctuations in the estimate of [the acceptance ratio]…”

Last Friday, in Paris-Dauphine, my PhD student Changye Wu showed me a paper of Ceperley and Dewing entitled the penalty method for random walks with uncertain energies. Of which I was unaware of (and which alas pre-dated a recent advance made by Changye).  Despite its physics connections, the paper is actually about estimating a Metropolis-Hastings acceptance ratio and correcting the Metropolis-Hastings move for this estimation. While there is no generic solution to this problem, assuming that the logarithm of the acceptance ratio estimate is Gaussian around the true log acceptance ratio (and hence unbiased) leads to a log-normal correction for the acceptance probability.

“Unfortunately there is a serious complication: the variance needed in the noise penalty is also unknown.”

Even when the Gaussian assumption is acceptable, there is a further issue with this approach, namely that it also depends on an unknown variance term. And replacing it with an estimate induces further bias. So it may be that this method has not met with many followers because of those two penalising factors. Despite precluding the pseudo-marginal approach of Mark Beaumont (2003) by a few years, with the later estimating separately numerator and denominator in the Metropolis-Hastings acceptance ratio. And hence being applicable in a much wider collection of cases. Although I wonder if some generic approaches like path sampling or the exchange algorithm could be applied on a generic basis… [I just realised the title could be confusing in relation with the current football competition!]