515 CANCELED: Machine Learning for Small, Uncertain, and Sparse Datasets
The world’s most valuable resource is no longer oil. It is data. Mining geoscience data is a key new research and application area in our field. Recently, a series of novel machine learning (ML) tools have been developed to analyze diverse datasets related to earth sciences, including climatic, geologic, geophysics, geothermal, carbon storage, oil/gas, and wildfire applications. Importantly, geologic datasets are often sparse or have missing values. Also, they are often small; therefore, data-hungry ML (e.g., deep learning) methods cannot provide accurate predictions. This workshop will talk about appropriate ML tools for handling sparse, uncertain, and small datasets. Specially, we will demonstrate various unsupervised and physics-informed unsupervised ML methods. Moreover, the workshop will also discuss the pros and cons of alternative machine learning methods for different real-world problems. The workshop will include seminar talks and hands-on real-time demonstrations of existing case studies. The workshop will be suitable and valuable for anyone regardless of their machine learning experience by providing materials at introduction, intermediate, and expert levels.
Saturday, 8 October 2022
8:00 AM - 5:00 PM
Presentations
No presentations found for this session.