Behavioural manifestations of epileptic seizures (ESs) and certain non-epileptic seizures (psychogenic non-epileptic seizures, or PNESs) have considerable overlap, and so discerning between these solely based on clinical criteria is difficult. Video EEG (electroencephalogram) monitoring (VEM) has high resource demands and is also expensive. We endeavour to classify seizures based on non-invasive measures.
Honours and Masters project
Displaying 91 - 100 of 243 honours projects.
Where does my electricity go?
Climate change will affect us all, and we have to do everything we can to minimize the magnitude of change. Investments in renewable generation help to reduce the impact of energy usage on the supply side, but that will not get us all the way there, especially in the near term. Consumers will also have to become much more efficient with their energy use.
Machine learning for comparing energy appliance usage across different demographics
Using relevant available data-sets, we compare appliance usage across households of different demographics. We then use machine learning techniques to infer how different households use different appliances at different times, resulting in diverse energy consumption behaviours.
Left/Right brain, human motor control and implications for robotics
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.
Left/Right brain, the hippocampus and episodic learning in AI/ML
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.
Left/Right brain in an RL agent
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.
Continual Few-shot reinforcement learning
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.
Automated Video-based Epilepsy Seizure Classification and Sudden Unexpected Death in Epilepsy (SUDEP) Detection
Develop, implement, and test deep learning techniques for automatic classification of epileptic seizures using video data of seizures
Ambulance Clinical Record Information Complexity
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.
Learning from massive amounts of EEG data
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].