Abstract

Speaker 1 – John Bronskill Title – VERSA : A Versatile and Efficient Framework for Few-shot Learning Abstract – VERSA is a new, probabilistic, amortized, multi-task, meta-learning framework. It achieves state-of-the-art performance in one-shot learning on the Omniglot and miniImageNet datasets, and produces compelling results on a one-shot ShapeNet view reconstruction challenge. This talk will give an overview of the system, describe how it works, and present results on few-shot classification and view reconstruction tasks. Speaker 2 – Rianne Van den Berg Title – Sylvester Normalizing Flows for Variational Inference Abstract – Variational inference relies on flexible approximate posterior distributions. Normalizing flows provide a general recipe to construct flexible variational posteriors. We introduce Sylvester normalizing flows, which can be seen as a generalization of planar flows. Sylvester normalizing flows remove the well-known single-unit bottleneck from planar flows, making a single transformation much more flexible. We compare the performance of Sylvester normalizing flows against planar flows and inverse autoregressive flows and demonstrate that they compare favorably on several datasets.

Joint work with Leonard Hasenclever, Jakub M Tomczak, Max Welling, to be presented at UAI 2018 . https://arxiv.org/abs/1803.05649