CTMTDS Newsletter 22 - November 2023

IN THE SPOTLIGHT 

Congratulations to Research Fellow Eden Li on being named a finalist in the 2023 Research and Engagement Awards. 

These awards recognise the achievements of Curtin researchers across all disciplines and stages of their career and acknowledges those who are making a positive contribution to raising the University’s research profile.  Eden presented her research virtually in the early career researchers section in partnerships.

Congratulations to Research Fellow Sirui Li for being named a ACS WA 1962 Prize and 1962 Medal Awards - finalist.

This prestigious award, sponsored by Professor Dennis Moore AM, showcases the best and brightest minds in WA and are a celebration of local talent, student excellence and the next generation of ICT professionals.

Hoa Bui was an invited speaker of the Continuous Optimisation cluster at the XL National Congress of Statistics and Operations Research and the XIV Conference on Public Statistics from 7 to 10 November 2023 in Elche, Spain. She shared her recent research on "Single-projection procedure for solving convex optimisation problems in Hilbert space."  She also gave two seminal talks at the Universität der Bundeswehr München (Munich, Germany) and the University of Alicante (Alicante, Spain). 

The AustMS WIMSIG Cheryl E. Praeger Travel Award funded her visit to the University of Alicante to collaborate with Professor Marco Lopez and Professor Abderrahim Hantoute, world-renowned experts in semi-infinite optimisation. They are collaborating to extend Hoa's recent results on single projection procedure to semi-infinite programming.

Hoa at the conference: XL National Congress of Statistics and Operations Research and the XIV Conference on Public Statistics

Melinda Hodkiewicz  along with Caitlin Woods  c

I

Michael Stewart  Large Language Models for Failure Mode Classification: An Investigation

Large Language Models (LLMs) have received a surge of interest recently due to their ability to encapsulate knowledge from a variety of domains. However, research has yet to explore the applicability of LLMs to maintenance. In this presentation, Michael provided details on his investigation into using LLMs to perform Failure Mode Classification (FMC), a critical maintenance task that reduces the need for reliability engineers to spend their time manually analysing work orders.

 Michael demonstrated his prompt engineering approach that enables fine-tuning an LLM (GPT-3.5) to predict a nominal failure mode code from a given observation. He illustrated that the LLM-based model outperforms a state-of-the-art text classification model, but only when the LLM is fine-tuned. Michael's investigation reinforces the need for high quality fine-tuning datasets for maintenance-specific tasks using LLMs.

Yingying Yang  - Long-term Integrated Maintenance Scheduling Optimization

An integrated mining site is a complex production system comprising many processing assets, intermediate buffer storages and linking facilities. Disruption of any unit may impact upstream/downstream product flow, stock levels, throughput, and the ability to satisfy customer demands. Thus, efficient coordination and strategic maintenance scheduling are essential to ensure a smooth flow of products and maximize the throughput.

Yingying presented a case study that explores the optimization of the interplay between maintenance timings and the whole system's performance over a long-term horizon. She  also gave a sensitivity analysis, adjusting parameters to gain more practical insights, such as how gaps between major shutdowns affect throughput and the key factors affecting overstock.

Ziyu Zhao presents Cy-Spider: Semantic Parsing Corpus and Baseline Models for a Property Graph Australasian Joint Conference on Artificial Intelligence held in Brisbane 2023.  

Melinda Hodkiewicz  along with Caitlin Woods present IDO maintenance reference ontology to  Tornbjorn Holm from Eurostep in Sweden and Erik Molin from SEIIA. They are planning to continue to collaborate in this area of maintenance ontologies and extracting information from unstructured maintenance texts.