SC²S Colloquium - October 05, 2017
|Date:||October 5, 2017|
|Time:||5:00 pm, s.t.|
Vincent Bode: Parallelization of a Sparse Grids Batch Classifier
The parallelization of a sparse grids batch classifier is evaluated, implemented, and analyzed. Several potential methods of distributing the workload are identified and evaluated for their feasibility, including class-based, spatial, and batch-based methods. From this analysis, a parallelization scheme that distributes work by batches of training data is chosen due to the flexibility it offers for less than ideal datasets, even though it involves more overhead. A design for the overall structure of the program, that makes use of a master/worker role distribution, and a communication plan are created and subsequently implemented. The semantics of the implementation are shown using software architectural techniques and several challenges are presented. Next, a guide for tuning the performance of the new learner is introduced, showing which variables need to be adjusted to decrease the time to solution on multi-node systems. Using strong and weak scaling tests, the implementation was then tested for consistency and efficiency. Results that scale almost linearly, or in one case super-linearly, were observed when comparing the old sequential to the new parallel implementation.
Irving Cabrera: The Use of Boosting Methods for Mineral Prospectivity Mapping within the ArcGIS Platform
Mineral prospectivity is the chance or likelihood that mineral deposits of the type sought can be found in a given area. To obtain this, geophysical, geochemical and geological information is combined to produce a prospectivity map. Methods like weights of evidence, logistic regression, neural networks, support vector machines and others have been used for this purpose. Methods of the boosting family are machine learning methods that combine simple classifiers to create a robust classification model. They are not yet used in fields prospectivity mapping. Several boosting algorithms were implemented inside Python tools to produce mineral prospectivity maps from real exploration data. These models are then evaluated in accuracy and robustness against modern methods.