Filter by faculty:

Found 24 matching student topics

Displaying 13–24 of 24 results

Data reasoning to extend domain knowledge in deep learning

A wide variety of companies now use personalized prediction models to improve customer satisfaction, for example, detecting cancer relapses, Detecting Attacks in Networks (e.g., SDN) or understanding Customer Online Shopping Behaviour. However, the dramatic increase in size and complexity of newly generated data from various sources is creating a number of challenges for domain experts to make personalized prediction.For example, early detection of cancer can drastically improve the chance and successful treatment. Recently, supervised deep learning has brought breakthroughs in …

Study level
PhD, Master of Philosophy, Honours, Vacation research experience scheme
Faculty
Faculty of Science
School
School of Computer Science
Research centre(s)
Centre for Data Science

When does computer vision fail?

Computer vision models predict where objects are in an image, and what those objects are. These vision models give robots the ability to perceive their environment and choose safe and smart actions based on this perception.Computer vision models can fail silently when exposed to unexpected or difficult environments - e.g. changes in camera viewpoints, changes in lighting, or when seeing new objects that haven't been seen before. This raises concerns about the safety of using vision models in the real …

Study level
Honours, Vacation research experience scheme
Faculty
Faculty of Engineering
School
School of Electrical Engineering and Robotics

Predicting player performance from one format to another in cricket

Identifying talent as early as possible in elite sport is critical. An important component of this is learning about what metrics of performance in lower grades to focus on to help predict performance in the top grade. This project will explore for this research problem for cricket.

Study level
Master of Philosophy, Honours, Vacation research experience scheme
Faculty
Faculty of Science
School
School of Mathematical Sciences
Research centre(s)
Centre for Data Science

Equation learning for partial differential equation models of stochastic random walk models

Random walk models are often used to represent the motion of biological cells. These models are convenient because they allow us to capture randomness and variability. However, these approaches can be computationally demanding for large populations.One way to overcome the computational limitation of using random walk models is to take a continuum limit description, which can efficiently provide insight into the underlying transport phenomena.While many continuum limit descriptions for homogeneous random walk models are available, continuum limit descriptions for heterogeneous …

Study level
PhD, Master of Philosophy, Honours, Vacation research experience scheme
Faculty
Faculty of Science
School
School of Mathematical Sciences
Research centre(s)
Centre for Data Science

Where’s the confusion? And can we make sense of it?

Confusion matrices characterise the performance of classification systems on training and test data, but they can be hard to make sense of, especially when there are many possible classes to which an example could be assigned.We have developed a new method to visualise confusion matrices and make distinct the contribution of the classifier and the contribution of the prior abundance of different classes.HypothesesIn situations where there is a suggestion that a classifier is biased, what insights can we gain by …

Study level
Vacation research experience scheme
Faculty
Faculty of Science
School
School of Computer Science
Research centre(s)
Centre for Data Science

Using machine learning to understand how the world’s microbiomes are changing due to climate

Shotgun metagenomic sequencing has become commonplace when studying microbial communities and their relationship with the health of our planet, and their direct effects on our own health. Currently, there are >180,000 shotgun metagenomes publicly available, but until recently trying to treat these data as a resource has been challenging due to its extreme size (>700 trillion base pairs).Recently we have developed a tool that can efficiently convert this base pair information into a straightforward assessment of which microorganisms are present …

Study level
Honours
Faculty
Faculty of Health
School
School of Biomedical Sciences
Research centre(s)

Centre for Microbiome Research

Quantifying oxygen tolerance in a climate-relevant ecosystem using machine learning

Aims and methodologyPermafrost thaw induced by climate change is predicted to make up to 174 Pg of near-surface carbon (less than 3m below the surface) available for microbial degradation by 2100. Despite having major implications for human health, prediction of the magnitude of carbon loss as carbon dioxide (CO2) or methane (CH4) is hampered by our limited knowledge of microbial metabolism of organic matter in these environments. Genome-centric meta-omic analysis of microbial communities provides the necessary information to examine how …

Study level
Vacation research experience scheme
Faculty
Faculty of Health
School
School of Biomedical Sciences
Research centre(s)

Centre for Microbiome Research

Evaluation of machine learning approaches for transfer learning

Transfer learning is becoming a popular machine learning approach which aims to transfer knowledge from a source domain to a target domain. Domain adaptation is a special case of transfer learning. Domain adaptation has proven to be significant in classification tasks where the target domain does not have labelled data, a requirement for building classifiers, however, there exists a related labelled data, called as source domain.Domain adaptation has been studied in the recent time and hence there exists many variants …

Study level
Vacation research experience scheme
Faculty
Faculty of Science
School
School of Computer Science
Research centre(s)
Centre for Data Science

Making machine learning understandable for human decision support in credit risk assessment

Credit risk assessment is an important aspect in personal loan applications because of an increase in loan applications and a decline in risk appetite of lending institutions. Traditionally, predictions of credit risk (i.e. the weighted probability of loan default) are based on individual socio-economic information, loan application information and dynamic transaction data. Although machine learning models have been applied to the evaluation of credit risk, there is a gap to their use for decision making because many of these methods …

Study level
Master of Philosophy
Faculty
Faculty of Business and Law
School
School of Accountancy

Explainability of machine learning methods

Machine learning can be very powerful but its black box reputation can be an obstacle for industry.In this project we'll look at creative ways to visually convey how methods like deep neural nets, random forest, XGBoost algoritms, and support vector machines work for different audiences.

Study level
Vacation research experience scheme
Faculty
Faculty of Science
School
School of Mathematical Sciences
Research centre(s)
Centre for Data Science

Quality and validation of machine learning methods

Not all machine learning methods are created equal and not all training datasets are of the same standard of quality.In this project we'll look at how machine learning methods can be quality assured and validated with a particular focus on training datasets.

Study level
Vacation research experience scheme
Faculty
Faculty of Science
School
School of Mathematical Sciences
Research centre(s)
Centre for Data Science

Human-in-the-loop techniques to debug machine learning models

Machine learning models are being deployed in critical domains such as healthcare, education and fintech. The current approach to deploying machine learning models is based on considering a data-centric approach where the models are evaluated using performance measures on a test set. However, the high performance of the model on test data is not indicative of its reliability,An important aspect of reliability is in the understanding of what exactly a machine learning model encodes, and to verify if it learns …

Study level
PhD, Master of Philosophy, Honours
Faculty
Faculty of Science
School
School of Information Systems

Page 2 of 2