Search

Recently, Deep Neural Network (DNN) algorithms have been explored for predicting trends in time series data. In many real world applications, time series data are captured from dynamic systems. DNN models must provide stable performance when they are updated and retrained as new observations becomes available. In this work we explore the use of automatic machine learning techniques to automate the algorithm selection and hyperparameter optimisation process for trend prediction. We demonstrate how a recent AutoML tool, specifically the HpBandSter framework, can be effectively used to automate DNN model development. Our AutoML experiments found optimal configurations that produced models that compared well against the average performance and stability levels of configurations found during the manual experiments across four data sets., Comment: arXiv admin note: text overlap with arXiv:2009.07943

Recently, a hybrid Deep Neural Network (DNN) algorithm, TreNet was proposed for predicting trends in time series data. While TreNet was shown to have superior performance for trend prediction to other DNN and traditional ML approaches, the validation method used did not take into account the sequential nature of time series data sets and did not deal with model update. In this research we replicated the TreNet experiments on the same data sets using a walk-forward validation method and tested our optimal model over multiple independent runs to evaluate model stability. We compared the performance of the hybrid TreNet algorithm, on four data sets to vanilla DNN algorithms that take in point data, and also to traditional ML algorithms. We found that in general TreNet still performs better than the vanilla DNN models, but not on all data sets as reported in the original TreNet study. This study highlights the importance of using an appropriate validation method and evaluating model stability for evaluating and developing machine learning models for trend prediction in time series data.

Current theories suggest that on Earth and, possibly, on other terrestrial planets early in their history, the first continental crust may has been produced by direct melting of hydrated peridotite. However, the conditions, mechanisms and necessary ingredients of such production remain elusive. To fill this gap, we have conducted experiments of serpentinite melting in the presence of variable proportions of basaltic melt, at typical conditions of the shallow lithosphere and asthenosphere. These experiments revealed formation of silica-rich liquids, which are similar to tonalite-trondhjemite-granodiorite magmas (TTG) identified in modern terrestrial oceanic mantle settings. Our results suggest a new mechanism of aqueous fluid-assisted partial melting of peridotite that may have operated on the early Earth and Mars just after the solidification of an ultramafic-mafic magma ocean, leading to the formation of the first embryos of continental crust. The proposed mechanism of the continental crust formation may have been predominant before the onset of plate tectonics., Comment: 62 pages, 9 figures, submitted to Frontiers in Earth Science

In response to a request by the Alfred P. Sloan Foundation to examine what attracted women to physics and mathematics, this report has two major goals: (1) to provide a literature review of published studies on women at the college and post college levels in science and mathematics; and (2) to present synopses of three studies carried out at the University of Michigan. The literature review is separated into five topics: (1) statistical trends about women in mathematics and physics; enrollment and graduation trends: decline in interest, and undergraduate programs producing students who earn advanced degrees; (2) enrollment and attrition studies investigating factors influencing women to pursue mathematics and science majors; (3) background characteristics of women in science, mathematics, and engineering; (4) determinants of women's achievement in mathematics and science (individual, environmental, and multiple factors); and (5) interventions employed to recruit and retain women in nontraditional fields of study. The three studies focused on what men and women in mathematics and physics courses believed enhanced and inhibited their decisions to pursue or not to pursue academic work and careers in mathematics and physics. The report ends with recommendations for institutional changes which would have the greatest impact on improving the academic experience of women in the fields of mathematics and physics. (Contains over 300 references.) (MDH)

The article analyses the development pathways to 2040 for Djibouti, Eritrea, Ethiopia, Somalia, Sudan and South Sudan using the International Futures modelling platform. Topics discussed include governance and security in the Horn of Africa countries, poverty and income inequality, and impact of the integrated development.