Archive for Uppsala University

SMC 2020 [en Madrid]

Posted in pictures, Statistics, Travel, University life with tags , , , , , , , , , , , on January 30, 2020 by xi'an

Palacio Real from Casa del Campo, on Nov. 10, 2011, during a pleasant bike ride through the outskirts of Madrid and along the renovated banks of Rio ManzanaresAn announcement for the incoming SMC 2020 workshop, taking place in Madrid next 27-29 of May! The previous workshops were in Paris in 2015 (at ENSAE-CREST) and Uppsala in 2017.  This workshop is organised by my friends Víctor Elvira and Joaquín Míguez. With very affordable registration fees and an open call for posters. Here are the invited speakers (so far):

Deniz Akyildiz (University of Warwick)
Christophe Andrieu (University of Bristol)
Nicolas Chopin (ENSAE-CREST)
Dan Crisan (Imperial College London)
Jana de Wiljes (University of Potsdam)
Pierre Del Moral (INRIA)
Petar M. Djuric (Stony Brook University)
Randal Douc (TELECOM SudParis)
Arnaud Doucet (University of Oxford)
Ajay Jasra (National University of Singapore)
Nikolas Kantas (Imperial College London)
Simon Maskell (University of Liverpool)
Lawrence Murray (Uber AI)
François Septier (Université Bretagne Sud)
Sumeetpal Singh (University of Cambridge)
Arno Solin (Aalto University)
Matti Vihola (University of Jyväskylä)
Anna Wigren (Uppsala University)

postdocs positions in Uppsala in computational stats for machine learning

Posted in Kids, pictures, Statistics, Travel, University life with tags , , , , , , , , , , , on October 22, 2017 by xi'an

Lawrence Murray sent me a call for two postdoc positions in computational statistics and machine learning. In Uppsala, Sweden. With deadline November 17. Definitely attractive for a fresh PhD! Here are some of the contemplated themes:

(1) Developing efficient Bayesian inference algorithms for large-scale latent variable models in data rich scenarios.

(2) Finding ways of systematically combining different inference techniques, such as variational inference, sequential Monte Carlo, and deep inference networks, resulting in new methodology that can reap the benefits of these different approaches.

(3) Developing efficient black-box inference algorithms specifically targeted at inference in probabilistic programs. This line of research may include implementation of the new methods in the probabilistic programming language Birch, currently under development at the department.

Sequential Monte Carlo workshop in Uppsala

Posted in pictures, Statistics, Travel, University life with tags , , , , on May 12, 2017 by xi'an

A workshop on sequential Monte Carlo will take place in Uppsala, Sweeden, on August 30 – September 1, 2017. Involving 21 major players in the field. It follows SMC 2015 that took place at CREST and was organised by Nicolas Chopin. Furthermore, this workshop is preceded by a week-long PhD level course. (The above picture serves as background for the announcement and was taken by Lawrence Murray, whose multiple talents include photography.)

efficient approximate Bayesian inference for models with intractable likelihood

Posted in Books, pictures, Statistics, University life with tags , , , , , , , , , , , , on July 6, 2015 by xi'an

Awalé board on my garden table, March 15, 2013Dalhin, Villani [Mattias, not Cédric] and Schön arXived a paper this week with the above title. The type of intractable likelihood they consider is a non-linear state-space (HMM) model and the SMC-ABC they propose is based on an optimised Laplace approximation. That is, replacing the posterior distribution on the parameter θ with a normal distribution obtained by a Taylor expansion of the log-likelihood. There is no obvious solution for deriving this approximation in the case of intractable likelihood functions and the authors make use of a Bayesian optimisation technique called Gaussian process optimisation (GPO). Meaning that the Laplace approximation is the Laplace approximation of a surrogate log-posterior. GPO is a Bayesian numerical method in the spirit of the probabilistic numerics discussed on the ‘Og a few weeks ago. In the current setting, this means iterating three steps

  1. derive an approximation of the log-posterior ξ at the current θ using SMC-ABC
  2. construct a surrogate log-posterior by a Gaussian process using the past (ξ,θ)’s
  3. determine the next value of θ

In the first step, a standard particle filter cannot be used to approximate the observed log-posterior at θ because the conditional density of observed given latent is intractable. The solution is to use ABC for the HMM model, in the spirit of many papers by Ajay Jasra and co-authors. However, I find the construction of the substitute model allowing for a particle filter very obscure… (A side effect of the heat wave?!) I can spot a noisy ABC feature in equation (7), but am at a loss as to how the reparameterisation by the transform τ is compatible with the observed-given-latent conditional being unavailable: if the pair (x,v) at time t has a closed form expression, so does (x,y), at least on principle, since y is a deterministic transform of (x,v). Another thing I do not catch is why having a particle filter available prevent the use of a pMCMC approximation.

The second step constructs a Gaussian process posterior on the log-likelihood, with Gaussian errors on the ξ’s. The Gaussian process mean is chosen as zero, while the covariance function is a Matérn function. With hyperparameters that are estimated by maximum likelihood estimators (based on the argument that the marginal likelihood is available in closed form). Turning the approach into an empirical Bayes version.

The next design point in the sequence of θ’s is the argument of the maximum of a certain acquisition function, which is chosen here as a sort of maximum regret associated with the posterior predictive associated with the Gaussian process. With possible jittering. At this stage, it reminded me of the Gaussian process approach proposed by Michael Gutmann in his NIPS poster last year.

Overall, the method is just too convoluted for me to assess its worth and efficiency without a practical implementation to… practice upon, for which I do not have time! Hence I would welcome any comment from readers having attempted such implementations. I also wonder at the lack of link with Simon Wood‘s Gaussian approximation that appeared in Nature (2010) and was well-discussed in the Read Paper of Fearnhead and Prangle (2012).