Fork me on GitHub

Swedish symposium on deep learning

Clarke error grid analysis on blood glucose predictions from our paper.

September 5-6, 2018, RISE AI is presenting work at the Swedish symposium on deep learning.

Poster: Blood glucose prediction with confidence

Welcome to come talk to us at our poster about blood glucose prediction. Our work was presented at the Blood glucose level prediction challenge at ICML this year and contains work on recurrent neural networks trained to predict the future blood glucose levels for type 1 diabetics with an approach that models the future probabilistically with an estimation of the standard deviation of the predictions. Joint work with John Martinsson (RISE AI), Alexander Schliep (GU), Björn Eliasson (Sahlgrenska university hospital), Christian Meijner (Chalmers), and Simon Persson (Chalmers).

Thursday, Sept 6: Character-based recurrent neural networks for morphological relational reasoning

Oral presentation about our work on morphological relational reasoning. Using recurrent neural network, and extra encoders to decide how to perform transformations, we attack the problem of inflecting words by seeing analogous word transformations. The presentation takes place between the keynotes by Chris Dyer and Joakim Nivre.

Master students wanted for exciting thesis projects in machine learning

Are you looking to write a master thesis in machine learning? Do you have a solid background on machine learning, neural networks, statistics, mathematics, and algorithms? Then you may be the one we are looking for to work on one of a number of projects that we are working on. In all the projects, we are performing research on state-of-the art techniques in machine learning, related to natural language processing, computer vision, and health applications.

Does this sound interesting? Do not hesitate to contact us. olof@mogren.one or +46-722-363990.

Swedish symposium on deep learning, 2018-09-05
Olof Mogren

Olof Mogren, PhD.