Research projects in Information Technology

Adversarial Machine Learning (AML) is a technique to fool a machine learning model through malicious input. Due to its significance in many scenarios, including security, privacy, and health application, AML has attracted a large amount of attention in recent years. However, the underlying theoretical foundation for AML still remains unclear and how to design effective and efficient attack and defence algorithms are remain a challenge in the research community. Furthermore, most existing  AML algorithms can only apply to Euclidean space.

The brain is a complex system and monitoring and imaging methods to observe critical neurophysiological variables underlying brain function are limited. This project works at the intersection of statistical signal processing, inference, machine learning and dynamical systems theory to develop new semi-analtyical filtering approaches for state and parameter estimation to infer neurophysiological variables such as network connection strengths between neural population networks underlying brain activity.

Motivation: In today’s increasingly digitalised world, software defects are enormously expensive. In 2018, the Consortium for IT Software Quality reported that software defects cost the global economy $2.84 trillion dollars and affected more than 4 billion people. The average annual cost of software defects on Australian businesses is A$29 billion per year.

When designing user interfaces, developers want to be able to position objects in a structured way such that controls are clear and neat.  In the past, absolute positioning and grids have been used for this purpose.  However, such rigid layout doesn’t now allow adaptive layout for interfaces that run on a variety of screen sizes or that need different control sizes due to application being internationalised into foreign languages.  For this reason, Apple introduced constraint-based GUI layout under the name Auto Layout for iOS developers.  With…

Digital versions of technical documents, like textbooks and academic papers, are usually produced as static PDF files. Research has shown that working with these on electronic devices is frustrating and inefficient, partly because people do not read such documents in a linear fashion as they do novels.

This PhD project will investigate how existing book contents (text and graphics/animations) can be automatically translated into accessible eBook formats with minimum human intervention. It will be part of the Books for the Vision Impaired and the GraVVITAS frameworks (www.monash.edu/it/inclusive-tech). The project will employ computer vision, image processing and human computer interaction techniques. It may also include hardware development of wearable assistive devices that use audio and haptic feedback. 

I am seeking PhD candidates interested in working on designing Learning Analytics innovations to study classroom proxemics by analysing and visualising indoor positioning data (along with other sources of evidence such as audio, physiological activity and characteristics of the students).

In this project, we are interested to translate speech to text. This can be used for translating a speaker who speaks in French to English in order to make it understandable by an English speaking person.

There are various aspects to this project, including (but not limited to): (i) handling multi-modal speech and text data, and (ii) producing simultaneous translations. This project involves innovating interesting models to address different aspects of this problem (eg [1,2]).

The meaning of an utterance depends on the broader context in which it appears. The context may refer to the paragraph, document, conversational history, or the author who has generated the utterance. In this project, we develop effective methods for translating text using the context, e.g. the rest of the sentences in the document or the conversational history.

I am seeking PhD candidates interested in working on designing Learning Analytics or similar reflection interfaces that automatically highlight design elements of data visualisations and generate narrative to communicate insights (instead of just plotting data).