DESCRIPTION:

The power of knowledge discovery has been emphasised with the advent of generative artificial (“Gen AI”) and the rapid uptake in language-focussed platforms like Open AI’s ChatGPT, Google “Bard” and Meta’s “Llama” models. The potential for knowledge discovery has since prompted much discussion on the potential for use in geoscience to address some of our greatest challenges, in particular the formation and discovery of mineral deposits that supply metals supporting the transition to a low-carbon economy. While these popular GenAI platforms are efficient at producing text-based responses to specific prompts, it is using these responses for modelling and simulation that is needed to support explorers. To do this, “semantic links” and the use of geoscience-focussed ontologies offer a framework to combined geological knowledge modelling with computational methods to address some of geosciences greatest challenges.

Work recently published with a code base demonstrates a petroleum thermal maturation and migration in the Ekofisk field, North Sea (Qu et al 2024). We want to take this approach and apply it with a mineral exploration focus on a mineral deposit-forming system. This work is linked to research being conducted with the MinEx CRC into developing a mineral systems-controlled vocabulary and ontology. The student will have access to subject matter expertise to assist with the geology and minerals concepts and understanding.

Tasks to be performed

  • Test code (Qu et al 2024 – available on GitHub)
  • Review of available and relevant geoscience ontologies
  • Support SMEs in coding up a conceptual modelling of geological knowledge representing a simple mineral-forming system.
  • Addition of “process triggers” relevant to the system (processes triggered when geological conditions are met)
  • Computational modelling
  • Visualisation of processes

 

We will apply the method to a mineral-forming system. From Qu et al 2024.

Background capabilities

Computer science, coding and NLP/KG experience essential. Knowledge of SPARQL, SHACL and RDF/OWL languages, or ability to learn desirable. Willingness to engage with subject matter expertise essential.

This is a fantastic opportunity to work with a multidisciplinary team of people on a real project of value. We are looking for someone who works well in a team environment and wants to leverage their coding and computer science skills as well as conduct novel research for their honours year. Conceptual models, ontologies and knowledge graphs are of increasing interest to the government and business communities

 

 

PROJECT NUMBER: 25/010

PROJECT TITLE: Structuring domain knowledge and semantic links to support mineral exploration

CLIENT: CSIRO

DISCIPLINE: Bachelor of Advanced Computer Science (Hons). Students with a WAM of 74 or higher.

SUPERVISORS: Prof. Melinda Hodkiewicz and Caitlin Woods

PROJECT TYPE: Full  (includes 8 weeks of onsite vacation work )

STUDENTSHIP: $10,000

DEADLINE: Sunday 1st December, 2024 via https://ceed.wa.edu.au/prospective-scholars/apply-now/