WA Tech Night - What is that anomaly? Using machine learning to obtain geological knowledge from downhole petrophysical data

Date = Wednesday 9 October

Time = 530 to 7pm

Venue = Celtic Club, West Perth WA

Title = What is that anomaly? Using machine learning to obtain geological knowledge from downhole petrophysical data

Speaker = Mark Lindsay, Senior Research Fellow, CET, UWA

Abstract:

Machine learning encapsulates methods that enable computers to learn and act as humans, while continuously improving their learning over time in an autonomous manner. The process of continuous improvement is a key advantage of machine learning, and is achieved via access to larger datasets and information. Unsupervised machine learning methods are particularly adept at analysing large and diverse datasets to find commonalities between attributes and produce 'clusters', or groups of data points that share similar characteristics. These techniques appear adaptable for geoscientific data, with uptake by practitioners being high in recent years. A geophysical study of the Eastern Yilgarn, Yamarna region, was supported by analysis of drillcore using machine learning, and attempted to obtain more geological knowledge from the measured data. Five drill cores are analysed in the M476 Project; 12DHDD0001, 12DHDD0002, 15EIS001, 15SYDD0003B and 15SYDD0004. Analyses produced some expected associations between rocks and petrophysics (high density and basalt) while some unexpected associations were also obtained (conductivity and quartz). These and other results will be presented, and examples of how these can be used to support structural geophysical interpretation will be discussed. This work was supported by the Mineral Research Institute of Western Australia, the Geological Survey of Western Australia and Gold Road Resources.

Speaker Bio:

Mark Lindsay is a Senior Research Fellow at the Centre for Exploration Targeting, School of Earth Sciences, the University of Western Australia (UWA) and specialises in structural geophysical interpretation, integrated geoscientific and 3D modelling with an interest in understanding their interrelated uncertainties and the value-of-information. He also has research interests that include investigating complex systems and mineralisation. Machine learning and AI are also disciplines Mark attempts to use to answer geological questions. Mark is co-leader of the Automated 3D modelling model of the 'MinEx' Co-operative Research Centre, a science leader in the Loop 3D geological modelling consortium.