Search

The ability to understand and model proteins’ dynamics is of great relevance in biology and medicine. A good comprehension of the way the proteins change their structure is important as these transitions give the function of the protein within the organism. In this paper we are introducing an unsupervised learning based approach using variational autoencoders, for uncovering protein motions and conformational transitions. The main goal of the research is to offer an interpretable method for proteins’ trajectories visualisation by learning a low dimensional space that accurately represents the input data, as empirically confirmed through the performed experiments. An additional aim is to comparatively evaluate the impact of two protein representations on the learning process.

Managing natural or human-made crisis events involves making decisions based on information gathered directly by first responders in the field using information and communication technologies (ICT), in combination with remote sensing (satellite, Earth Observation - EO) technologies delivering images and sensor data. For efficient crisis and disaster management, this information needs to be timely available and as accurate as possible, meaning that the available technologies need to be intelligently selected and combined to meet such requirements. In this short paper, we present such an approach in which semi-automatized analytics using combination of very high-resolution EO imagery and crowdsourcing technologies is demonstrated to improve and optimize the efficiency of crisis management before, during and after disaster events.