Mining Automation: A magazine-style article on the use of AI, robotics, machine learning and optimization in mining automation. This survey provides some insights on state of the art technologies and future trends; including mine rehabilitation and low-carbon energy transition.

Hyperspectral CNN: This work describes a self-supervised convolutional neural network that classifies mineral groups in hyperspectral images. It considers issues such as shadows and illumination invariance. It uses a stacked auto-encoder, data (spectral relighting) augmentation and transfer learning.

Geostatistics: These works are concerned with using robust estimators to (a) measure geochemical consensus within a group of samples which requires extending the breakdown point of the MCD estimator; (b) removing statistical outliers from compositional data.

Spatial Algorithms: A Bayesian method for rectifying inaccuracies in an existing surface given new geochemical observations in the form of bench blast-hole assays. Estimating geological domain likelihood using machine learning.

Computer Graphics: Modelling orebody structures using variational methods. Updating block structure given mesh surfaces of geological boundaries. 

Rock Mechanics: Proposed method detects coal-seam transitions based on a change in the rotation-to-thrust power ratio. The cross-section reveals multiple coal seams in the ground using measure-while-drilling techniques.

Signal processing: Image compression - This conference paper examines the performance gaps that exist when the Karhunen–Loève transform (KLT) is applied in different ways to exploit redundant information along the slide axis within a chest CT scan (medical volumetric image). It also shows a motion-compensated wavelet transform (implemented in a "lifting" framework) can essentially achieve the theoretical gain.

Signal processing: Video coding and perceptual quality - This work proposes a method for reducing wavelet compression artefacts in scalable video coding based on visual masking principles.