We are pleased to announce that there have been five (5) successful Queensland recipients in Round 2 of the Artificial Intelligence for Decision Making Initiative (AI4DM).
Each recipient will receive three months funding of up to $20,000 per project. This funding, from numerous Commonwealth agencies, is administered on a state basis.
Recipients could also have the chance to join a national network that concentrates on developing AI (Artificial Intelligence) and ML (Machine Learning) technology.
Project prototypes created because of the funding may also be considered for progression through the Defence Artificial Intelligence Centre (DAIC) or defence and university sectors.
Meet the 5 Recipients
1. Dr Abigail Koay
Research Fellow in Cyber Security, University of Queensland – alongside team member, Dr Naipeng Dong.
Project title: A Lightweight and Adaptive LSTM Model in Large-Scale Dynamic Environment
Snapshot of project description: “This project aims to address these limitations by developing a new lightweight and adaptive method to update deep learning models automatically only when concept drift is detected.”
Potential outcome: “We anticipate that this project will show the potential of the lightweight adaptive LSTM (long short-term memory) approach in retaining past knowledge and reducing of cost of retraining the LSTM model. This project will act as the first stage of building an ideal cyber-attack detector that can be used to enhance cyber security defences to protect national critical infrastructure.
2. Ranju Mandal
Postdoctoral Researcher at CQ University
Project title: Designing a Novel Deep Learning-based Application for Automatic Document Parsing
Snapshot of project description: “This project aims to design a novel prototype deep learning-based architecture for parsing and extracting information from diverse unstructured or semi-structured documents such as manual, communications, user-created pdf, etc., and store in a structured form for further analytics.”
Potential outcome: “The system will be a secure, searchable, central repository for records revealing insights via analytics tools. It will help users digitally archive, track, control life cycle (e.g. legal contracts), reduce cost, and search documents in a faster, efficient, and interactive way while reserving them securely.”
3. Associate Professor Marcus Gallagher
Artificial Intelligence Group in the School of Information Technology and Electrical Engineering, University of Queensland
Project title: Solving Realistic Portfolio Optimisation Problems Using Interactive Multi-objective Evolutionary Algorithms
Snapshot of project description: “The project proposed is to investigate the formulation of the portfolio design problem as a multi-objective problem and the application of interactive multi-objective evolutionary algorithms to solving instances of these problems.”
Potential outcome: “The project will develop a prototype software tool that can be used to provide good solutions to portfolio investment problems using interactive multi-objective evolutionary algorithms. This should establish proof-of-concept of the approach. The project would provide a clear indication of the feasibility of developing a production-quality tool based on the techniques involved for future use by Defence.”
4. Dr Kenneth Li-minn Ang
Professor of Electrical and Computer Engineering, University of the Sunshine Coast
Project title: Multi-Modal Narrative Extraction
Snapshot of project description: “Our proposed research addresses an open challenge to be able to extract useful knowledge from a multimodal collection of data items such as text, audio, images and video. We propose an approach to utilise concept graph learning (CGL) models”
5. Dr Zhe Wang
Senior Lecturer at the School of ICT and a researcher in artificial intelligence (AI), Griffith University
Project title: Enhancing Semantic Data Integration with Rule Learning and Reasoning
Snapshot of project description: “In this project, we plan to develop effective methods to handle uncertainty in temporal knowledge graphs. In particular, to perform rule learning and reasoning with quadruplets that are associated with confidence degrees, in a way that the uncertainty in knowledge graphs can be effectively utilized to measure the quality of the rule models and enhance the accuracy of link prediction”.
Potential outcome: “This project aims to enhance the capability of Defence and National Security in the semantic integration of uncertain and evolving data. It is expected to lead to technologies that will be useful for to construct and enrich their knowledge graphs. The efficiency of the system will allow online validation of hypotheses for decision making, and the explainability of the rule-based model will offer insights over higher-order event patterns as well as transparency in the decision process.”
QDSA Upcoming Events and Opportunities
Want to find out more about the QDSA’s upcoming events and opportunities?
Check out our News section on our website here or subscribe to our e-newsletter here.
QDSA Upcoming Events and Opportunities
QDSA has many upcoming events and opportunities, with more being added each week. Want to find out what is on the horizon? Check out our News section on our website here or subscribe to our e-newsletter here.
The Queensland Defence Science Alliance (QDSA) is a university-led initiative to grow and connect an agile Defence innovation ecosystem, leveraging Queensland’s strengths, to deliver trusted solutions to meet Defence requirements.