Below you can find descriptions for the four lighting talks at AHM21 on January 27.
How maps are made
Ståle Haug, The Norwegian Mapping Authority
The process of making maps has changed significantly over the last decades. Ståle Haug of the Norwegian Mapping Authority shows us how maps are made today using the latest technology.
How deep neural networks understand human language
Andrey Kutuzov, researcher, Language Technology Group (LTG), University of Oslo
Andrey Kutuzov from UiO's Language Technology Group will describe modern natural language understanding technologies based on machine learning and word co-occurrence statistics in large collections of texts. These architectures (ELMo, BERT, GPT-3, etc.) are all trained on the simple task of language modeling, but still, yield surprisingly good results, sometimes even considered a break-through. Kutuzov will also present recently trained language understanding models for Norwegian.
Pathology? Digital?? And artificially intelligent???
Joel Hedlund, Center for Medical Image Science and Visualization
(CMIV), Linköping University
The convergence of advanced imaging, automation, and powerful analytics in healthcare is bringing together the tools needed for scientists and clinicians to leverage new developments in medical science. Joel Hedlund will explain the current state-of-the-art and provide a glimpse of the digital future of pathology.
Weather nowcast - when will it rain?
Christoffer Elo, The Norwegian Meteorological Institute
The Norwegian Meteorological Institute uses a fast optical flow scheme to produce a weather nowcast, processing large amounts of radar data every fifth minute. Christoffer Elo will talk about the method behind the 90 minutes nowcast and give examples of dos and don'ts on how to run a smooth production.
