Event Details

It is part of regular Statistics seminar series hosted by the statistical learning group. All are welcome - in particular the topic is in between statistical learning and sequential Monte Carlo.

Title: Adversarial Sequential Monte Carlo

Speaker: Kira Kempinska, UCLAbstract: How can we perform efficient inference in directed probabilistic models with intractable posterior distributions? In this talk, I will introduce a new technique for improving finite-sample inference approximations by learning highly flexible proposal distributions for sequential importance samplers, such as particle filters. We represent proposal distributions as implicit generative models, that is models that you can sample from but which do not have an explicit parametric form, and train them using variational inference rephrased as a two-player game, hence establishing a principled connection between Sequential Monte Carlo (SMC) and Generative Adversarial Networks (GANs). Our approach achieves state-of-the-art performance on synthetic and real inference problems. The development of the method has been motivated by the need to infer true vehicle positions from noisy GPS sequences, but is equally applicable to other inference problems with intractable posterior distribution.Background reading: - Kempinska, K., Shawe-Taylor, J.,"Adversarial sequential Monte Carlo"- Mescheder, L., et al., "Adversarial Variational Bayes"- Paige, B., Wood, F., "Inference networks for sequential Monte Carlo in graphical models"