Honours and Minor Thesis projects

In this project, you will build an autonomous agent in the MineRL environment for playing Minecraft or an agent for Animal-AI.  Herein, you will learn how to incorporate symbolic prior knowledge for improving the performance of an agent trained by using deep reinforcement learning (RL) technique, which is the core technique to build AlphaGo.

The existing deep learning-based time series classification (TSC) algorithms have some success in multivariate time series, their accuracy is not high when we apply them to brain EEG time series (65-70%). This is because there is a large variation between EEG data of different subjects, so a TSC model cannot generalise on unseen subjects well. In this research project, we investigate self-supervised contrastive learning to encode the EEG data. This way we can better model the distribution of our EEG data before classifying it into different mental statuses. See recent work here [1].

Turning Point’s National Ambulance Surveillance System is a surveillance database comprising enriched ambulance clinical data relating to alcohol and other substance use, suicidal and self-injurious thoughts and behaviours, and mental health-related harms in the Australian population. These data are used to inform policy and intervention design and are the subject of ever-increasing demand from academic professionals and units, government departments, and non-government organisations.

Develop, implement, and test deep learning techniques for automatic classification of epileptic seizures using video data of seizures

This project takes a different approach to RL, inspired by evidence that Hippocampus replays to the frontal cortex directly. It is likely used for model building, as opposed to the mainstream view in cognitive science and ML - where 'experience replay' ultimately improves policy. The predicted benefits are sample efficiency, better ability to generalize to new tasks and an ability to learn new tasks without forgetting old ones.

The brains of all bilaterally symmetric animals on Earth are divided into left and right hemispheres. The anatomy and functionality of the hemispheres have a large degree of overlap, but they specialize to possess different attributes. This principle is poorly understood and has not been exploited in AI/ML. The right hemisphere is more dominant for novelty, and the left for routine. Activity slowly moves to the left hemisphere as a task is perfected. In this project, we apply that principle to continual RL, where new tasks are introduced over time…

The hippocampus is critical for episodic memory, a key component of intelligence, and a sense of self. There are a number of computational models, but none of them consider the fact that the hippocampus is, like the rest of the brain, divided into Left and Right hemispheres. Division into Left and Right is poorly understood, but undoubtedly critical, as it is a remarkably conserved feature of all bilaterally symmetric animals on Earth.

The brains of all bilaterally symmetric animals on Earth are divided into left and right hemispheres. The anatomy and functionality of the hemispheres have a large degree of overlap, but they specialize to possess different attributes. This principle is poorly understood and has not been exploited in AI/ML. Previously, we mimicked biological differences between hemispheres, and achieved specialization and superior performance in a classification task that matched behavioral observations.

Are you interested in applying your database knowledge to a real project? This project aims to develop a patient registry for hospitals around Australia. This is a collaboration with the Faculty of Medicine, Monash University. We will be building a central database or a data warehouse repository to store patient admission to the hospitals.

Are you interested in applying your AI/DL knowledge to the medical domain?