Honours and Minor Thesis projects

A major challenge in cancer therapeutics is to kill tumour cells without harming normal cells in the body. Traditional chemotherapy tries to do this by killing cells that are fast dividing, a characteristic hallmark of cancer cells, however as many other cells in the body are also fast dividing – such as those in the hair and the gut – chemotherapy typically results in undesirable side effects. Newer targeted therapies are designed to specifically target cancer cells, by exploiting the genetic changes that distinguish tumour cells from normal cells.…

Despite enormous progress in research, cancer remains a devastating disease worldwide. Since generally not all patients will respond to a specific therapy, a great challenge in cancer treatment is the ability to predict which patients would benefit (or not) to a therapy of choice. This helps improve treatment efficacy and minimise unnecessary sufferings by non-responders. There is thus a pressing need to identify robust biomarkers (i.e. genes/proteins) that can accurately predict the right patients for the right drugs.…

Antimicrobial resistance (AMR) continues to evolve as a major threat to human health and new strategies are required for the treatment of AMR infections. Bacteriophages (phages) that kill bacterial pathogens are being identified for use in phage therapies, with the intention to apply these bactericidal viruses directly into the infection sites in bespoke phage cocktails. Using such a biological agent for infection control requires deep understanding of the phage.

Multidrug resistance (MDR) poses critical challenges to global health. In 2017 the World Health Organization identified Acinetobacter baumannii, Pseudomonas aeruginosa, and Enterobacteriaceae as the top-priority pathogens that urgently require development of novel therapeutic options. Recently, bacteriophage therapy has attracted extensive attention owing to its potential of being used as novel antimicrobials to combat MDR pathogens.

Neural module networks (NMNs) [1] support explainable question answering over text [2] by parsing a natural-language question into a program. Such a program consists of a number of differentiable neural modules that can be executed on text in a soft way, operating over attention scores. As a result, NMNs learn to jointly program and execute these programs in an end-to-end way.

The performance of deep neural models rely on large amounts of labeled data, however, most data remain unlabeled in the real world scenario. While annotating data is expensive and time consuming, active learning seeks to choose the most appropriate and worthwhile data for human annotation. It is noticed that humans give labels to some specific data with some labeling reasons or rationales,  which are often existing in the data.  The goal of this research is to develop effective deep active learning techniques with rationales.…

Mission-critical systems have to comply to various formal standards – e.g. DO-178C and ISO26262 - about their operation, usually heavily relying on formal specification languages such as TLA+. This presents many challenges to developers in terms of how to write, read and communicate the target system’s formal specifications. In most cases, having the right formal methods experts to write specifications does not solve the problem as the wider development team needs to be able to deeply understand the formal specifications.

User reviews on app distribution platforms such as Google Play store and Apple App store are a valuable source of information, ideas, and requests from users. They reflect the needs and challenges users encounter including bugs, feature requests, and design. Recent research has shown that reviews can also serve as a proxy for understanding the values of the users and how users perceive that their values have been violated by the mobile app/mobile app developers. However, there are limited studies that show whether mobile app updates fix violations of the user's values and to what extent.…

This project will develop new technologies for supervised machine learning from time series building upon our world-leading and award winning research in the area. See my time series research for details of the research program on which this research will build.

The world is dynamic, in constant flux. However, machine learning typically learns static models from historical data. As the world changes, these models decline in performance, sometimes catastrophically so. This PhD will develop technologies for addressing this serious problem, building upon our groundbreaking research into the problem.