Educating for earthquake science and risk in a tectonically slowly deforming region

Over the past decade, scientists have been called to participate more actively in public education and outreach (E&O). This is particularly true in fields of significant societal impact, such as earthquake science. Local earthquake risk culture plays a role in the way that the public engages in educational efforts. In this article, we describe an adapted E&O program for earthquake science and risk. The program is tailored for a region of slow tectonic deformation, where large earthquakes are extreme events that occur with long return periods. The adapted program has two main goals: (1) to increase the awareness and preparedness of the population to earthquake and related risks (tsunami, liquefaction, fires, etc.), and (2) to increase the quality of earthquake science education, so as to attract talented students to geosciences. Our integrated program relies on activities tuned for different population groups who have different interests and abilities, namely young children, teenagers, young adults, and professionals.

Text classification using compression-based dissimilarity measures

Arguably, the most difficult task in text classification is to choose an appropriate set of features that allows machine learning algorithms to provide accurate classification. Most state-of-the-art techniques for this task involve careful feature engineering and a pre-processing stage, which may be too expensive in the emerging context of massive collections of electronic texts. In this paper, we propose efficient methods for text classification based on information-theoretic dissimilarity measures, which are used to define dissimilarity-based representations. These methods dispense with any feature design or engineering, by mapping texts into a feature space using universal dissimilarity measures; in this space, classical classifiers (e.g. nearest neighbor or support vector machines) can then be used. The reported experimental evaluation of the proposed methods, on sentiment polarity analysis and authorship attribution problems, reveals that it approximates, sometimes even outperforms previous state-of-the-art techniques, despite being much simpler, in the sense that they do not require any text pre-processing or feature engineering.