Bayesian Neural Networks for Geothermal Resource Assessment: Prediction with Uncertainty

We consider the application of machine learning to the evaluation of geothermal resource potential. A supervised learning problem is defined where maps of 10 geological and geophysical features within the state of Nevada, USA are used to define geothermal potential across a broad region. We have available a relatively small set of positive training sites (known resources or active power plants) and negative training sites (known drill sites with unsuitable geothermal conditions) and use these to constrain and optimize artificial neural networks for this classification task. The main objective is to predict the geothermal resource potential at unknown sites within a large geographic area where the defining features are known. These predictions could be used to target promising areas for further detailed investigations. We describe the evolution of our work from defining a specific neural network architecture to training and optimization trials. Upon analysis we expose the inevitable problems of model variability and resulting prediction uncertainty. Finally, to address these problems we apply the concept of Bayesian neural networks, a heuristic approach to regularization in network training, and make use of the practical interpretation of the formal uncertainty measures they provide.

The Effect of Ground Truth Accuracy on the Evaluation of Localization Systems

The ability to accurately evaluate the performance of location determination systems is crucial for many applications. Typically, the performance of such systems is obtained by comparing ground truth locations with estimated locations. However, these ground truth locations are usually obtained by clicking on a map or using other worldwide available technologies like GPS. This introduces ground truth errors that are due to the marking process, map distortions, or inherent GPS inaccuracy. In this paper, we present a theoretical framework for analyzing the effect of ground truth errors on the evaluation of localization systems. Based on that, we design two algorithms for computing the real algorithmic error from the validation error and marking/map ground truth errors, respectively. We further establish bounds on different performance metrics. Validation of our theoretical assumptions and analysis using real data collected in a typical environment shows the ability of our theoretical framework to correct the estimated error of a localization algorithm in the presence of ground truth errors. Specifically, our marking error algorithm matches the real error CDF within 4%, and our map error algorithm provides a more accurate estimate of the median/tail error by 150%/72% when the map is shifted by 6m.