Join us for another thought-provoking event with Shuyan from Central South University, China. In this EarthByte Seminar Series event, Shuyan Yu will talk about “Novel Machine Learning Approaches for Identifying Multivariate Geochemical Anomalies Considering Spatial Structures.”.

Seminar Details:

• Date: Wednesday, 08 May 2024
• Time: 11 am – 12 pm AEDT
• Location: Geoscience Conference Room 449, online 
• Zoom Link: https://us05web.zoom.us/j/86264411015?pwd=7JEiBdVphaJV1KxGHErmwjVr1BPuKO.1&from=addon

Novel Machine Learning Approaches for Identifying Multivariate Geochemical Anomalies Considering Spatial Structures

Machine learning methods have shown success in Geochemical anomaly identification critical in mineral exploration. However, fully leveraging the inherent spatial-elemental structures of geochemical data remains challenging. We proposed two novel geochemical anomaly recognition methods that utilize tensor dictionary learning (machine learning) and Transformer models (deep learning), respectively, effectively exploring spatial and elemental structure in geochemical data. Tensor dictionary learning method can leverage tensor representations to capture spatial and elemental structures in geochemical data, facilitating anomaly identification. Transformer model incorporates self-attention mechanisms to capture non-local spatial dependencies, offering an end-to-end framework for anomaly detection. Through case studies in the northwest Jiaodong Peninsula, China, the effectiveness of both approaches in identifying geochemical anomalies were validated. Both studies showcase the potential of machine learning techniques to advance mineral exploration efforts.