Výsledky vyhledávání - "Thin, Achille"

Report

BR-SNIS: Bias Reduced Self-Normalized Importance Sampling

Autor: Cardoso, Gabriel, Samsonov, Sergey, Thin, Achille, Moulines, Eric, Olsson, Jimmy

Importance Sampling (IS) is a method for approximating expectations under a target distribution using independent samples from a proposal distribution and the associated importance weights. In many applications, the target distribution is known only

Externí odkaz: http://arxiv.org/abs/2207.06364

Zobrazit plný text záznamu

Report

Monte Carlo Variational Auto-Encoders

Autor: Thin, Achille, Kotelevskii, Nikita, Doucet, Arnaud, Durmus, Alain, Moulines, Eric, Panov, Maxim

Variational auto-encoders (VAE) are popular deep latent variable models which are trained by maximizing an Evidence Lower Bound (ELBO). To obtain tighter ELBO and hence better variational approximations, it has been proposed to use importance samplin

Externí odkaz: http://arxiv.org/abs/2106.15921

Zobrazit plný text záznamu

Report

NEO: Non Equilibrium Sampling on the Orbit of a Deterministic Transform

Autor: Thin, Achille, Janati, Yazid, Corff, Sylvain Le, Ollion, Charles, Doucet, Arnaud, Durmus, Alain, Moulines, Eric, Robert, Christian

Sampling from a complex distribution $\pi$ and approximating its intractable normalizing constant Z are challenging problems. In this paper, a novel family of importance samplers (IS) and Markov chain Monte Carlo (MCMC) samplers is derived. Given an

Externí odkaz: http://arxiv.org/abs/2103.10943

Zobrazit plný text záznamu

Report

Nonreversible MCMC from conditional invertible transforms: a complete recipe with convergence guarantees

Autor: Thin, Achille, Kotelevskii, Nikita, Andrieu, Christophe, Durmus, Alain, Moulines, Eric, Panov, Maxim

Markov Chain Monte Carlo (MCMC) is a class of algorithms to sample complex and high-dimensional probability distributions. The Metropolis-Hastings (MH) algorithm, the workhorse of MCMC, provides a simple recipe to construct reversible Markov kernels.

Externí odkaz: http://arxiv.org/abs/2012.15550

Zobrazit plný text záznamu

Report

MetFlow: A New Efficient Method for Bridging the Gap between Markov Chain Monte Carlo and Variational Inference

Autor: Thin, Achille, Kotelevskii, Nikita, Denain, Jean-Stanislas, Grinsztajn, Leo, Durmus, Alain, Panov, Maxim, Moulines, Eric

In this contribution, we propose a new computationally efficient method to combine Variational Inference (VI) with Markov Chain Monte Carlo (MCMC). This approach can be used with generic MCMC kernels, but is especially well suited to \textit{MetFlow}

Externí odkaz: http://arxiv.org/abs/2002.12253

Zobrazit plný text záznamu