Since 2015, Solve Geosolutions has been trusted by the minerals exploration and mining industry to provide specialist geoscientific data science and high-powered computational geophysics and geoscience advice and services.
As Australia’s first dedicated geoscientific machine learning consultancy, we have seen the market react to and adopt new technologies to improve new and existing workflows that make the most of the rapid increase in the volume and diversity of geoscientific data.
We need a new generation of tools to harness the complexity of this data and extract everything we can from our dataset. At Solve, our aim is to develop and implement data science-based solutions that improve how we mine and explore large and complex multivariate datasets.
We believe that education in the burgeoning field of geoscientific ML is the key to bringing the right technologies to the right processes, driving our industry forward.
Our consultants are from diverse backgrounds and specialise in both geoscience, data science, mathematics and mining, which means our data science workflows are grounded in a strong understanding of geology.
After studying structural geology and geophysics at Monash University, Brenton has consulted as a structural geophysicist for PGN Geoscience as well as working in a variety of geological and geophysical roles—predominantly in exploration. Most recently, Brenton has worked as a geophysicist and data scientist for MMG Exploration on a range of commodities. Brenton has a passion for the integration of regional geophysical and remote sensing datasets using machine learning and prospectivity analysis.
The technical lead on most of Solve’s projects, Tom is an experienced mathematical geologist with a strong background in statistics, computer science and geology. Tom develops, integrates and implements fit-for-purpose machine learning solutions that enable clients to expand their knowledge base. Outside work hours, Tom can be found at sports grounds around Melbourne either playing (badly) or supporting (far too enthusiastically) any sport that involves people chasing a ball.
Liam is a geoscientist based in Melbourne and specialises in geophysics, 3D geological modelling (Leapfrog) and the integration of multivariate datasets in 3D. Originally from Tasmania, Liam studied at the University of Tasmania where he majored and completed his honours in geophysics, focusing on machine learning methods in blast hole ore categorisation. From greenfields exploration through to project mine geology, Liam has worked across a variety of commodities and terranes. Outside of geology, Liam is an avid sports fan and enjoys playing, watching and using machine learning to win fantasy sports.
Mark is an established geoscientist with significant experience driving and developing innovation within global mineral exploration programs. His background in geophysics and mathematics has seen him work in a wide range of settings, from developing algorithms in physics laboratories to exploring the Andean copper belt.
He hopes to promote the acceptance of machine learning and computer vision in the mining and exploration industries by building practical, effective solutions to geoscientific problems.
Yasin joined Solve after a career in mine and blast engineering.
He transitioned into geostats and data science via PhD and post-doc positions, both locally and abroad.
At Solve, he provides key insight into the geotechnical aspects of how data science techniques can increase geological and geotechnical understanding.
Harvey came to Solve from a computer science and full-stack development background. He transitioned into data science and more specifically computer-vision through a Master’s degree.
At Solve, he implements cutting-edge computer vision technologies to geoscientific imagery, including core photos, remote sensing data, and other high-resolution datasets.
Tom is a geophysicist with a keen interest in applied geophysics. He graduated with first class honours from the University of Tasmania in 2012 and spent several years working as a field geophysicist specialising in potential field survey methods.
After gaining extensive experience in the field, he returned to Tasmania to commence work on a PhD project investigating the 3D electrical structure of the Tasmanian lithosphere using magnetotelluric methods.
Tom recently completed his PhD and is applying his new-found programming and computing skills to traditional geophysical consulting as well as data science projects.
Luisa is an experienced geophysicist with significant experience in the operation of large multi-geophysical field programs within global mineral exploration programs.
Her background in petrophysics and geophysical interpretation has seen her work in a wide range of commodities in exploration and mine sites across Australia, southern Africa and Eurasia.
Luisa has a passion for greenfields and brownfields exploration and promotes the integration of geoscientific datasets and data analytics for the smart, successful discovery of mineral deposits.
Amir did his bachelor’s degree in Mining Engineering, master’s and PhD in Petroleum Engineering – Exploration Geophysics. Throughout his career, he has worked in various roles ranging from geophysicist, data scientist, petrophysicist and well log engineer.
As a geophysicist and data scientist, he has a strong background in quantitative interpretation, signal processing and applications of computer vision and machine learning in seismic model building.
Solve uses a wide variety of mathematical, statistical, computational and machine-learning workflows to obtain geological insights for our mining and exploration clients.
This process generally begins with exploratory data analysis to understand the structure in the data. Once the problem is clearly understood, there are a number of solutions that Solve can provide, which are centred around the characterisation and prediction of geoscientific data.
Read on to learn about some of the broad areas in which we operate; exploration, orebody knowledge, traditional and machine-learning boosted geoscience.
Finding the next deposit, or extensions to existing deposits
Left: Underappreciated expression of mineralised BIF unit in the Pilbara, Australia. Right: Sn/W prospectivity in north-east Tasmania, Australia. Bottom: Multi-model prospectivity analysis showing overlap of disjointed models.
Greenfields terranes present explorers with a unique set of challenges, from collating and integrating public and proprietary disparate datasets, through to determining the best allocation of limited resources.
At Solve, we help explorers make the most out of all scales of data, from globally-available remote sensing products, to state or national geophysical compilations, through to drilling and surface geochemistry databases.
These data, along with terrane subject matter expertise enter our data modelling and machine learning workflows to provide products including mineral prospectivity analysis, refined data products and geophysical data modelling and inversions.
Our greenfields processes lead to refined geological mapping, improved camp-scale targeting, better understanding of large-scale geological structures, and can help direct drill programs.
Brownfields teams face different problems; their goal is to identify extensions to known deposits or find satellite deposits close to existing infrastructure.
Improved understanding of these extensions or satellite systems can be obtained by leveraging the knowledge of existing or near-by mineral systems.
At this geological scale, we typically operate in a more data-rich environment than in the greenfields setting. This enables us to focus on identification and characterisation of key data relationships that inform geological and mineral-systems understanding. This is typically an informed-modelling process, where we ask scientifically reasonable questions from our data.
Such informed modelling processes include geologically constrained inversion models, data gap analysis, proximity modelling and 3D geology model validation, resulting in improved resource and drill targeting, guidance of future data acquisition campaigns, and target ranking.
Near-mine, multi-scale, multi-data, analysis of orebody expression.
Understanding and modelling variation in deposit character
Resource definition, classification and estimation are critically important in determining mine planning and scheduling, whilst determining the financial viability and profitability of a project.
As a project progresses towards an advanced stage, there is a sharp increase in the quantity and variability of new data that drives characterisation of the variation in lithological, mineralogical, metallurgical and geotechnical character of the ore body.
Solve aids the process of improving orebody knowledge by integrating disparate datasets, building near-miss indication models, performing quantitative and qualitative machine-driven models, and assessing data inconsistencies.
The results of these processes are data and expert-driven orebody models and near real-time OBK insights. These are driven by productionised machine learning and computer vision (see Datarock) workflows that supercharge resource-intensive work and provide previously unobtainable, orebody-wide datasets.
Animated examples of several mine-scale models that improve orebody understanding.
The heart of exploration and mining
Our team has a strong geoscientific background, including significant experience across a range of geoscientific disciplines. We offer technical and project management services across the following:
Left: Delivered AGG data. Right: The result of Solve’s variably-isotropic noise-reduction filter. Zoom, pan and slide to explore.
Geoscientific principles and spatial context are at the core of all machine learning and advanced modelling that Solve undertakes.
In addition to employing existing state-of-the-art geoscientific and geospatial processing algorithms, we have developed several advanced routines to extract new insights from existing data.
These processes span the geophysical, remote sensing, analytical geochemistry and mineralogy disciplines. The goal is to provide the best-possible representation of data, facilitating improved geoscientific understanding.
An example of one such processing algorithm is shown, where a variably isotropic filter is applied to airborne gravity gradiometry data in order to minimise geologically irrelevant signal while retaining signal with maximum coherence along the localised prevailing geological strike.