Computer science

Supervisor: Dr. Rahul Gopinath

Eligibility: You will work with the supervisor directly for this project. Excellent skills in programming (Expertise networking protocols like HTTP etc), problem solving, as well as the ability to work independently are required.

Project Description:

Activity pub is a suite of specifications that allow social interaction similar to facebook and twitter (X), and is a federated solution that can be installed by the individuals. Activity pub provides an API for delivering notifications and subscribing to content.

The main issue that is holding back ActivityPub from taking off is that the user needs to obtain either an account from a central resource such as a Mastodon server or host an instance on their own in a web facing machine somewhere that can implement the ActivityPub protocol.

However, this may not be necessary. It may be possible to build a minimum viable activity pub instance that can provide a limited number of features such as distribution of blog content, liked and boosted posts, that can be hosted statically. Similarly, the users could consume the content meant for this instant either from a second tool hosted on a static webpage in the same domain, or a completely separate tool. This project asks the student to understand the ActivityPub protocol, and build such a minimum viable tool.

This project may be extended to an Honours thesis with the supervisor, and if completed successfully could be the basis of a publication at a top SE venue.

(In your application, please indicate if you are interested in a Honours application with the supervisor).

Requirement to be on campus: No

Supervisor: Dr. Rahul Gopinath

Eligibility: You will work with the supervisor directly for this project. You should be a fast learner. Excellent skills in Python, Javascript, HTML, problem solving, as well as the ability to work independently are required.

Project Description:

Computational Notebooks such as Jupyter notebooks have taken the world by storm. However, there are many features they lack including the ability to host the notebook inside a /single/ static HTML page without requiring network access that are required in many learning environments.

We have developed a static HTML page that implements the basic notebook features along with additions such as the ability to save sessions and restart at specific points. This project asks the student to implement several novel usability features into this notebook and make it on par with or beyond Jupyterlite feature wise.

This project may be extended to an Honours thesis with the supervisor, and if completed successfully could be the basis of a publication at a top SE venue.

(In your application, please indicate if you are interested in a Honours application with the supervisor).

Requirement to be on campus: No

Supervisor: Dr. Rahul Gopinath

Eligibility Criteria: You will work with the supervisor directly for this project. You should be a fast learner. Excellent skills in programming (Basic Python & Java knowledge is necessary), problem solving, as well as the ability to work independently are required.

Project Description:

In cybersecurity, your fuzzers can often come up with a very large input that causes a crash. To debug the crash, one often needs a much smaller input that reproduces the crash. HDD is an algorithm used to reduce the input size if the input conforms to a grammar. However, it does not work well if the grammar is incomplete such as in many handwritten parsers and the input can’t be parsed.

In this project, we will explore input-repair based techniques to repair and then parse the input and compare it with using original delta debugging (which does not require a grammar) and coverage-based strategies to identify hierarchies in the partially parsed input.

This project may be extended to an Honours thesis with the supervisor, and if completed successfully could be the basis of a publication at a top SE venue.

(In your application, please indicate if you are interested in a Honours application with the supervisor).

Requirement to be on campus: No

Supervisor: Dr. Rahul Gopinath

Eligibility: You will work with the supervisor directly for this project. Excellent skills in unix and programming, problem solving, as well as the ability to work independently are required.

Project Description:

Makefile the venerable UNIX build system has numerous problems, and several build systems have been proposed to overcome the problems of makefiles. One of the best known is called redo by Daniel J Bernstein, which fixes most of the problems in make. However, one of the problems with redo is that it is not well supported in legacy systems where programming languages such as Python are not available. Furthermore, a full implementation of redo requires filesystem support for listening to filesystem events, and identifying changed elements.

This project asks the student to implement the redo in posix AWK without external dependencies such as executables. Such a build system can finally be used as a replacement for make.

This project may be extended to an Honours thesis with the supervisor, and if completed successfully could be the basis of a publication at a top SE venue.

(In your application, please indicate if you are interested in a Honours application with the supervisor).

Requirement to be on campus: No

Supervisor: Dr. Rahul Gopinath

Eligibility: This project requires a fast learner, who is comfortable with doing literature search, understanding the algorithms, along with the implementation.

Project Description:

Data aggregation and cleaning is one of the most important steps in data science. The data may come from multiple sources and hence, may not match the required format exactly. This is especially an issue if the required format is a rich format like JSON, XML, S-Expr etc. In such cases, we may have to rely on available error correction algorithms (at best) or manual labour (at worst). While numerous error correction algorithms exist, error correction for context-free grammars is still lagging with the best-known algorithm from 1972 from Aho et al. which extends Earley parsing. However, the algorithmic complexity of the error correcting parser is O(n^3).

This project will involve (1) understanding the background by doing literature review, (2) understanding the algorithm from the research paper, (3) implementing the algorithm for error correction using Valiant’s recognizer (4) empirical analysis of this algorithm to identify its actual performance characteristics.

This project may be extended to an Honours thesis with the supervisor, and if completed successfully could be the basis of a publication at a top SE venue.

(In your application, please indicate if you are interested in a Honours application with the supervisor).

Requirement to be on campus: No

Supervisor: Dr. Rahul Gopinath

Eligibility: You will work with the supervisor directly for this project. You should be a fast learner. Excellent skills in programming (Python & Java knowledge is necessary), problem solving, as well as the ability to work independently are required.

Project Description:

For fuzzing, one needs the input specification. However, in many circumstances, such specifications are unavailable. The alternative here is to try and infer the specification from the given program under blackbox conditions. One of the best techniques known so far is called the TTT algorithm.

This project asks you to implement the TTT algorithm for blackbox grammar inferrence, and compare it with the L* algorithm (given).

This project may be extended to an Honours thesis with the supervisor, and if completed successfully could be the basis of a publication at a top SE venue.

(In your application, please indicate if you are interested in a Honours application with the supervisor).

Requirement to be on campus: No

Supervisor: Dr. Rahul Gopinath

Eligibility: You will work with the supervisor directly for this project. You should be a fast learner. Excellent skills in programming (Basic Python & Java knowledge is necessary), problem solving, as well as the ability to work independently are required.

Project Description:

Bugs are often induced by similar inputs. If we can identify and isolate such inputs, we can prevent bug inducing code from being exercised, and hence prevent vulnerabilities. The requirement for getting this to work is a large set of test cases. Fortunately, we know how generate a large number of test cases through fuzzing.

This project will explore how to learn regular expressions from such bug inducing tests such that any input that matches such regular expressions is guaranteed to trigger the bug. Such regular expressions can be then used to combine different bugs to produce more novel buggy behavior.

This project may be extended to an Honours thesis with the supervisor, and if completed successfully could be the basis of a publication at a top SE venue.

(In your application, please indicate if you are interested in a Honours application with the supervisor).

Requirement to be on campus: No

Supervisor: Dr. Rahul Gopinath

Eligibility: You will work with the supervisor directly for this project. You should be a fast learner. Excellent skills in functional programming with Haskell, problem solving, as well as the ability to work independently are required.

Project Description:

Mutation analysis is a technique used for evaluation of test suites and test generators. It involves generating syntactic variants of the program with tiny changes, and checking whether the test suite is able to detect the change in comparison to the original. In this project, you will work on a Haskell mutation analysis engine called MuCheck, updating the codebase to the latest, and use it to evaluate the testedness of Haskell projects.

This project may be extended to an Honours thesis with the supervisor, and if completed successfully could be the basis of a publication at a top SE venue.

(In your application, please indicate if you are interested in a Honours application with the supervisor).

Requirement to be on campus: No

Supervisor: Dr. Rahul Gopinath

Eligibility: This project requires a fast learner, who is comfortable with doing literature search, understanding the algorithms, along with the implementation.

Project Description:

Parsers are one of the key interfaces between computers and human beings. A general parser takes a specification (typically provided as a grammar) the input string and generates a derivation tree. A key difficulty when using parsers is that in many languages, specific operators have specific precedence. For example, multiplication (*) typically has higher precedence than addition (+). A typical way to account for this is to try to encode this precedence in the grammar itself. However, this is error-prone, and has led to a bad reputation for general context-free parsers. The alternative is to extract the first parse tree and transform the tree to the derivation tree with the correct precedence bracketing. LaLonde et al showed how to do this for simple languages (LaLonde 1981).

This project will involve (1) doing the background literature review (2) understanding and implementing LaLonde’s algorithm (3) applying this algorithm on common programming language patterns that typically are solved with handwritten precedence parsers and evaluating the effectiveness.

This project may be extended to an Honours thesis with the supervisor, and if completed successfully could be the basis of a publication at a top SE venue.

(In your application, please indicate if you are interested in a Honours application with the supervisor).

Requirement to be on campus: No

Supervisor: Dr. Rahul Gopinath

Eligibility: You will work with the supervisor directly for this project. You should be a fast learner. Excellent skills in programming (wizardry in Python is necessary), problem solving, as well as the ability to work independently are required.

Project Description:

Python is well known for having many syntactical warts such as its handling of default arguments to procedures, no multiline lambdas, etc. Many of these can be worked around, and Python provides a mechanism to do that called Codecs, which allow you to parse the Python file before it is interpreted. All you need is a single line in Python source file specifying the codec to be used.

This project asks the student to build a complete syntax for Python that avoids as much of the warts of Python as possible, parse such source files, and encode them into equivalent Python programs, allowing alternative syntax for Python.

This project may be extended to an Honours thesis with the supervisor, and if completed successfully could be the basis of a publication at a top SE venue.

(In your application, please indicate if you are interested in a Honours application with the supervisor).

Requirement to be on campus: No

Supervisor: Dr. Rahul Gopinath

Eligibility: You will work with the supervisor directly for this project. You should be a fast learner. Excellent skills in programming (Basic Python & Java knowledge is necessary), problem solving, as well as the ability to work independently are required.

Project Description:

For fuzzing, one needs the input specification. However, in many circumstances, such specifications are unavailable. The alternative here is to try and infer the specification from the given program under blackbox conditions. The current grammar inference systems are GLADE and Arwada from software engineering. They require a set of samples that cover all features of the program so that they can do grammar inference.

The traditional GI algorithms like L* RPNI and TTT also require a distribution of inputs. In this project, you will compare the effectiveness of random sampling vs prefix-based samples (see Fuzzing with Fast Failure Feedback) for L*, RPNI, Arvada and GLADE.

Requirement to be on campus: Yes *dependent on government’s health advice

Supervisor: Dr. Rahul Gopinath

Eligibility: You will work with the supervisor directly for this project. You should be a fast learner. Excellent skills in programming (Basic Python & Java knowledge is necessary), problem solving, as well as the ability to work independently are required.

Project Description:

Fuzzers routinely produce very large and incomprehensible inputs that can crash a program. Such inputs can't be debugged unless a much smaller input that reproduces the crash can be obtained.  Delta debugging is an algorithm used to reduce the input of a test case to make it comprehensible.

One of the problems with delta debugging is that while reducing the input, new inputs may be produced that induces an unrelated crash. This is called test reduction slippage. The problem is that slippage can result in programmers time being wasted on an unrelated problem. This project will explore how to solve the test reduction slippage using execution grammars and compare it to the effectiveness of using simple coverage.

This project may be extended to an Honours thesis with the supervisor, and if completed successfully could be the basis of a publication at a top SE venue.

(In your application, please indicate if you are interested in a Honours application with the supervisor).

Requirement to be on campus: No

Supervisor: Dr. Rahul Gopinath

Eligibility: You will work with the supervisor directly for this project. You should be a fast learner. Excellent skills in programming (Excellent skills in Python, Scheme and Lisp would be essential), problem solving, as well as the ability to work independently are required.

Project Description:

Python is often described as an acceptable Lisp. If this is the case, can we adapt the object system in lisp (Common Lisp Object System) which is known to be extremely powerful to Python? This project asks the student to implement and document the complete TinyClos (a variant of CLOS on the scheme language) to Python without using any of the advanced features of python including class-based inheritance.

This project may be extended to an Honours thesis with the supervisor, and if completed successfully could be the basis of a publication at a top SE venue.

(In your application, please indicate if you are interested in an Honours application with the supervisor).

Requirement to be on campus: No.

Supervisor: Dr. Rahul Gopinath

Eligibility: You will work with the supervisor directly for this project. You should be a fast learner. Excellent skills in programming (Excellent Python knowledge is necessary), problem solving, as well as the ability to work independently are required.

Project Description:

Mutation analysis is a technique used for evaluation of test suites and test generators. It involves generating syntactic variants of the program with tiny changes, and checking whether the test suite is able to detect the change in comparison to the original. In this project, you will develop a python-bytecode mutation analysis engine, and use it to evaluate the testedness of Python projects.

This project may be extended to an Honours thesis with the supervisor, and if completed successfully could be the basis of a publication at a top SE venue.

(In your application, please indicate if you are interested in a Honours application with the supervisor).

Requirement to be on campus: No

Supervisor: Prof Eduardo Velloso

Eligibility:  

Required

• Experience with C# development
• Interest in VR/MR/AR development

Desired

• Experience with Unity
• Experience with the Unity XR and Meta XR SDKs

Project Description:

We seek interns to contribute to the development of a mixed reality platform (Quest 3) to allow users to collaborate remotely across physically different environments, in large groups, and asynchronously.

Example features that you might work on include:

• Strategies for seamlessly aligning different physical spaces
• Text summarisation and speech generation for asynchronous meetings
• Managing large group meetings in MR (e.g. multiple users controlling the same avatar)
• 3D recordings of meetings

This project will give you experience in designing novel interactions for emerging technologies, mixed reality development in Unity, and continuous improvement based on user feedback.

Requirement to be on campus: Yes * dependent on government’s health advice.

Supervisor: A/Prof Joseph Lizier

Eligibility: Strong programming ability (most project options require Java, except experiments/demonstrations which require one of Python or Matlab). Strong mathematical analysis required for project option “incorporate new statistical enhancements”. The fundamentals of information theory can be learned on the project.

Project Description:

The JIDT toolkit provides a stand-alone, open-source code Java implementation (also usable in Matlab, Python, etc.) of information-theoretic measures of distributed information processing in complex systems. In other words, it is a data analysis tool to measure how information is stored, transferred and modified in natural systems like the brain.

There are a number of enhancements planned for JIDT to enhance the that students could contribute to, e.g.:

• Enhance the GUI to support more calculation types Re-engineer APIs for different types of estimators for better alignment
• Optimise memory use to handle larger data types
• Design experiments/demonstrations to test new estimators
• Incorporate new statistical enhancements

The enhancement(s) that a student contributes to in the project will be determined based on their interests and skills.

Requirement to be on campus: No.

Supervisor: A/Prof. Joseph Lizier

Eligibility: Strong numerical programming ability, particularly in Matlab; strong mathematical capability for the “extend”/”relate” options. Background in graph theory / complex networks or control theory would be useful.

Project Description:

Synchronisation phenomena – where the disparate units in a system collectively self-organise into a common rhythm – emerge in a wide variety of systems, from heart cells to neural activity to fireflies flashing. Recently we developed a new mathematical approach relating the underlying network structure in such systems (how the units are connected in a graph or complex network) to the ability of such networks to synchronise.

Opportunities exist for students to extend (mathematically-focussed) and/or experiment with (more computationally-focussed) such models in a range of scenarios and types of dynamics in order to establish the range of their applicability.

These include:

• Experimenting on the range of applicability for our analytic solution to the (non-linear) Kuramoto model;
• Conducting experiments on how coupling delay affects sync;
• Extend and experiment with how to design noise to enhance sync;
• Relate sync to other measures of complex networks such as “network communicability” in the computational neuroscience domain.

Requirement to be on campus:  No

Supervisor: A/Prof Lijun Chang

Eligibility: Good algorithm design and C (or C++) programming skills

Project Description:

We are nowadays facing a tremendous amount of large-scale social networks with millions or billions of edges. Thus, there is a need of designing efficient algorithms for processing large-scale graphs.

In this project, our aim is to design efficient algorithms to speed up graph processing on ever-growing large graph datasets. The problems that we will be investigating can be (1) dense subgraph (e.g., clique, near-clique) computation over a large sparse graph which finds one dense subgraph of the maximum size, or (2) dense subgraph enumeration which enumerates all maximal dense subgraphs.

Requirement to be on campus: No

Supervisor: Dr. Anusha Withana

Eligibility: You will work with the supervisor and a PhD student, and we expect you are a fast learner. Experience with 3D-printing and design skills are highly desirable.

Project Description:

The unpredictability of seizures can increase the risk of physical injury and cause mental stress for people living with epilepsy. Existing research has shown that it is possible to detect or predict seizures through collecting and processing biological signals. However, despite the potential benefits of these technologies, the device adoption rate and adherence rate remain low.

We want to develop a personalised wearable device that is able to cater to real user needs, in order to improve the quality of life of people with epilepsy.

Requirement to be on campus: Yes *dependent on government’s health advice

Supervisor: Dr. Anusha Withana

Eligibility: You will work with the supervisor and a PhD student, and we expect you are a fast learner. Excellent skills in programming, and human computer interaction are added benefits

Project Description:

What happens if a person has four hands?

Virtual Reality (VR) has altered the way we interact with digital environments, bringing a new era of immersive experiences in education, work, and well being. 

An emerging interaction paradigm in VR is shared control or embodiment -- a concept where several individuals or entities share a common virtual avatar or form. In this project, we concentrate on the concept of a single avatar with four hands to investigate novel user-avatar dynamics, introducing a potentially more versatile interaction method. The project will look into how two users work together to control one avatar that has four hands.

Requirement to be on campus: Yes *dependent on government’s health advice  

Supervisor: Dr. Anusha Withana

Eligibility: You will work with the supervisor and a PhD student, and we expect you are a fast learner. Excellent skills in programming, , machine learning, and human computer interaction are added benefits.

Project Description:

On-skin sensors are gaining traction in human-computer interaction field to facilitate easy and ergonomic interactions. However, designing on-skin sensors by prototyping could take a long time if we want to personalize them to each individual and optimize them to each application. Can we use machine learning to design the optimal on-skin sensor for each person and each application?

In this project we explore how machine learning algorithms can be used to generate sensor patterns under some given specifications, within a small amount of time, ideally in a few minutes. The project will build on our ongoing work about easily manufacturable on-skin sensor designs.

Requirement to be on campus: Yes *dependent on government’s health advice

Supervisors: Dr Anusha Withana and Dr Hossein Moeinzadeh

Eligibility: Background in Python programming is required. Having knowledge in machine learning and statistical analysis methods are encouraged.

Project Description: Dive into the world of water science with our cutting-edge project, where we unravel the mysteries of water quality using the power of prediction! Imagine being able to tell which water parameters are the most crucial for keeping our rivers, lakes, and drinking water safe and clean. That's exactly what we're aiming to do.

Using data analysis and machine learning, we'll sift through numerous factors—like temperature, pH levels, and contaminants—to pinpoint which ones truly matter, and find the relations between parameters. You'll get hands-on experience with real-world data, learn data science tricks, and contribute to a project that could make a big splash in environmental protection. Join us in making waves in water quality research!

You can find relative information on this project in our recent publication.

Requirement to be on campus: Yes * dependent on government’s health advice.

Supervisor: Prof Alan Fekete

Eligibility: Experience running applications on databases, and measuring performance properties of systems. Skill in deploying systems.

Project Description:

Measure impact of using different durability settings for updates running against a replicated MongoDB database. Measurements would be both performance and number of updates lost. The aim is to check against a theory developed by a former Honours student.

Requirement to be on campus: No

Supervisor: Prof Alan Fekete

Eligibility: Experience running applications on databases, and measuring performance properties of systems. Skill in deploying systems.

Project Description:

Measure impact of using different isolation levels for transactions running against a database (maybe PostgreSQL, maybe MongoDB). Measurements would be both performance and number of data integrity violations produced. For a suitable student, maybe designing a new benchmark (requires good theory and mathematical reasoning).

Requirement to be on campus: No.

Supervisor: Dr. Zhanna Sarsenbayeva

Eligibility: To do well in this project you will need statistic skills, programming skills, independent learning skills, independent decision making skills.

Project Description:

Virtual reality (VR) is a growing area in human-computer interaction. However, it is hard for people with short-sighted or farsighted to explore VR clearly without wearing glasses. This project aims to propose a method for people with short-sighted or farsighted to help them explore VR without wearing actual glasses or contact lenses.

The method is to move the scene forward/backward without changing the overall visual proportions. The joysticks on the controllers can be used to bring the scene forward/backward and zoom out/in at the same time. Besides, we can input a fixed number based on the user's contact lens as the pre-setting. The study can provide several types of experiments to test the effect of the method.

SCOPE: You will work with one supervisor and one PhD student, andwe expect you are a fast learner. We expect you have basic Python, R, C# skills, and familiar with Unity and Blender software. We will give you guidance to learn how to solve aresearch problem, and you will have a deep understanding of accessibility in virtual reality.

EXPECTED OUTCOME: A paper in one of the top venues of the HCI research field.

References: 1, 2, 3

Requirement to be on campus: Yes *dependent on government’s health advicee.

Supervisor: Dr. Zhanna Sarsenbayeva

Eligibility: To do well in this project you will need programming skills, independent learning skills, independent decision making skills.

Project Description:

Virtual reality (VR) is a growing area in human-computer interaction. In VR, interaction predominantly relies on the use of thumb and index finger pinching motions. This standard method, however, excludes individuals who lack these fingers.

This project aims to design an innovative interaction technique that requires only a single finger. This inclusive approach ensures that even those who have lost various fingers can still effectively interact using their hands.

This method will begin by allowing users to select their preferred finger, followed by enabling them to execute essential tasks such as clicking, moving, and dragging. The finger stays in an object after 3seconds, then we treat this object as they aim to select. Clicking can be done using double click. Moving and dragging can be done using dragging. Cancelling an object can be done through curling fingers.

SCOPE: You will work with one supervisor and one PhD student, andwe expect you are a fast learner. We expect you have basicPython, R, C# skills, and familiar with Unity and Blender software. We will give you guidance to learn how to solve a research problem, and you will have a deep understanding ofaccessibility in virtual reality.

EXPECTED OUTCOME: A paper in one of the top venues of the HCI research field.

References: 1, 2, 3

Requirement to be on campus: Yes *dependent on government’s health advice.

Supervisor: Associate Professor Simon Poon

Eligibility: Preference will be given to students with strong interests in multi-disciplinary research. Combined degree students are encouraged to apply. Background in Information Systems, Health Informatics, AI applications, Statistics, Social Sciences and are advantageous.

Project Description:

The idea of using technologies and specially the use of AI to enable technological progress for sustainable development goals has emerged as an important topic for research.  This project aims to apply a novel logical evidence synthesis to explore the complex causality of recent digital health progress on achieving the sustainable development goals (SDGs).

The intent of using the approaches in evidence synthesis (like systematic reviews, meta-analysis and set-theoretic approach) is to uncover that factors in complex configurations may play different roles as core and periphery factors in achieving SDGs. Certain important factors being overlooked in traditional empirical studies but they could play a non-trivial role in achieving the development targets.

The set-theoretic based framework may serve as both an analytical tool for understanding causality from complex interdependencies among provisions of digitalization of health, as well as for facilitating the abstraction of those complexities through means of pinpointing core and periphery within the bundle of seven indicator-categories captured in the Global Digital Health Index.

Requirement to be on campus: No

Supervisor: Prof Jinman Kim

Eligibility: Experiences with Deep Learning architecture

Project Description:

In the emerging field of vision–language pretraining (VLP), the advent of large language models (LLMs) for textual supervision has seen breakthroughs in guiding the computer vision models to learn novel visual representations for downstream applications, including in biomedical image domains.

The aim of the project is to develop a new VLP model by leveraging knowledge-enhanced image (vision) and text (language) derived semantics. We will use our existing semantic extraction algorithms and use its output as the knowledge. We will benchmark the developed algorithm on visual-query answering (VQA) downstream task.

Requirement to be on campus: Yes *dependent on government’s health advice

Supervisor: Prof. Athman Bouguettaya

Eligibility: Good Programming Skills; Data Handling and Management Skills; Experience on Artemis will be a plus.

Project Description:

Social media has evolved into an essential part of our lives. Social media users upload a lot of content on different social media platforms. The content shared on social media may contain untrustworthy images. Existing approaches relies on image processing to detect fake images.

These solutions are costly and computationally intensive. Some recent solutions investigate comments on a post to assess credibility of an image. These solutions may not accurately determine the credibility of an image because fake posts on social media can still get supportive comments. 

We leverage the metadata of an image and the related posted information to determine the trustworthiness of a crowdsourced image. This information may reflect some changes in the image. For instance, there may be some discrepancies in the metadata which might be an indication of modifications in the image. Therefore, investigating this meta-information may provide useful insights about the changes introduced in the image.

Requirement to be on campus: Yes *dependent on government’s health advice

Supervisor: Prof. Athman Bouguettaya

Eligibility: Good programming background in either Java or Python, and good knowledge on Algorithms.

Project Description:

The composition of crowdsourced IoT services poses several trust-related challenges. New composite services are created when a single service cannot satisfy the consumer's requirements. For instance, assume a crowdsourcing environment where IoT devices provide computing services. In such an environment, IoT service providers may offer computing resources (CPU, memory) to perform processing tasks for other IoT devices.

A computation resource-poor device like a smartwatch may use these resources to do computationally intensive tasks like rendering a map. However, a potential IoT service consumer may have concerns regarding the trustworthiness of the service providers.

A distrustful service provider might not protect the privacy of the consumers' data or provide unreliable performance. Similarly, an IoT service provider may have concerns regarding their consumers' trustworthiness. Malicious consumers can misuse IoT services by sending malicious software.

The project seeks to explore methodologies that can ensure the integrity of the crowdsourced IoT environment, mitigating potential risks associated with such misuse.

Requirement to be on campus: Yes *dependent on government’s health advice

Supervisor: Prof. Athman Bouguettaya

Eligibility: Good Programming Skills; Data Handling and Management Skills; Experience in drone flight simulators will be a plus.

Project Description:

A continuous expansion of urban areas is leading to an increased demand for instant deliveries from warehouses to customers' doorsteps. Unmanned Aerial Vehicles (UAVs) or drones have the potential to serve customers with timely and cost-effective deliveries. Drones usually operate in a skyway network which is an interconnected set of nodes.

The nodes are building rooftops which serve as recharging stations or delivery destination for drones. Drones may recharge at nodes for long mile deliveries as they are constrained by limited battery capacity. These nodes are connected through skyway segments. Multiple drones may share the same skyway segment at the same time to transit between nodes.

This project aims to use advanced AI techniques to predict and identify areas in a skyway network where drone aerodynamics could disrupt operations. The goal is to optimize drone efficiency by addressing potential interference zones proactively.

Requirement to be on campus: Yes *dependent on government’s health advice

Supervisor: Prof. Athman Bouguettaya

Eligibility: Good Programming Skills; Data Handling and Management Skills; Experience on Artemis will be a plus.

Project Description:
Digital communication platforms, ranging from Social Network Sites (SNSs) like Twitter to reputable news channels such as BBC, have emerged as the primary medium for information exchange. Images are the primary format through which a significant portion of information is shared. Recent studies indicate that a considerable amount of these images is manipulated to skew public perception.

The rise of generative-Artificial Intelligence (AI) techniques has led to the creation of AI-generated fake images, called deepfakes. Deepfakes pose a significant challenge to the integrity of information dissemination. Current approaches to detect deepfakes rely on computationally intensive image processing techniques.

Additionally, initiatives like the C2PA aim to authenticate images by incorporating icons for manual verification. However, these approaches present practical limitations given the vast volume of images circulated on these platforms daily.

To address this issue, we leverage meta-information inherent in AI-generated images. Through a thorough examination of these attributes, our objective is to cultivate valuable insights that can be applied effectively in the realm of deepfake detection.

Requirement to be on campus: Yes *dependent on government’s health advice

Supervisors: Dr. Zhanna Sarsenbayeva, Dr. Anusha Withana

Eligibility:

Project Description:

Smart wearables have become regular day-to-day equipments that help people easily manage and organise their lives. These are capable of identifying different emotional/mental states of the users based on different physical metrics including the heart-rate, electrodermal responses, and sweat levels.

Currently, various applications are available in smart devices to regulate the identified stress and anxiety levels of people by suggesting activities like meditation, breathing exercises, etc. The purpose of this project is to explore the possibility of integrating smart wearables with voice assistants, so that whenever an abnormal emotional state is detected, it will enquire and lets the user talk about it.

Parallel to the speech input, it will capture the context data using the sensors available in all connected devices (eg: - smartphone). After analysing the inputs, the voice assistant will suggest possible solutions, or instruct on possible actions to overcome the issue.

The differentiation from this application with the existing emotional regulation apps available in smart wearables is that it is able to capture the users’ speech to understand the exact situation they are in. Further, it can combine the speech input with the context to provide a more personalised solution than a one-size-fits-all solution.

SCOPE: Software development, smart wearables, smart systems

EXPECTED OUTCOME: An application which integrates smart wearables (smart watch) with a voice assistant to detect abnormal emotional states of the user and analyse the user’s speech input and context input to provide personalised solutions.

Requirement to be on campus: Yes *dependent on government’s health advice.

Supervisor: A/Prof Uwe Roehm

Eligibility: Strong academic background in Computer Science, especially databases and ideally also machine learning (the latter can be picked up during the project too). This project also requires good Python programming skills. Experience in benchmarking and tuning of systems would be desirable.

Project Description:

MongoDB is a popular NoSQL database system which allows to partition data distributed over multiple nodes for fast performance and scalability. An important design decision is how data gets indexed so that only specific partitions need to be contacted for a given query. In previous work, we have shown that MongoDB’s own query optimiser sometimes selects a sub-optimal query execution plan by wrongly estimating the benefit of indexes.

Together with MongoDB, we are working on an external query advisor that is using deep reinforcement learning (DRL) to direct MongoDB towards the optimal plan by monitoring past query executions and then adding suitable execution ‘hints’ to new queries. 

In this project, the task is to take our existing query advisor and benchmark its performance against the latest version of MongoDB. Most of the required code, especially the DRL part, already exists from a previous project, written in Python.

This project will give good insight into the usage of machine learning for the auto-tuning of database systems, as well as into MongoDB in particular – the project is a collaboration with MongoDB, and its outcome is planned for being used in a publication about our approach.

Requirement to be on campus: No

Supervisor: Dr. Xi Wu

Eligibility: Strong knowledge of Mathematics, especially Discrete Mathematics; Good at programming

Project Description:

As one of the most important applications of wireless sensor networks, Flying Ad Hoc Networks (FANETs) have been widely used in various situations. Due to the nature of node mobility in FANETs, the reliability of communication becomes a common problem which may result in abnormalities and decrease the quality of service provided by the wireless system.

In this project, we are going to use the process algebra and model checking tools to formally model and analyse the communication protocols used in the networks, whose results could guarantee the reliability and correctness of messages in the normal communication process. Some other interesting security related properties may also be verified in the project.

Requirement to be on campus: No

Supervisor: Dr. Sasha Rubin

Eligibility: At least DI in Models of Computation or Algorithm Design.

Project Description:

Planning is sub-area of AI which involves finding a finite-state program that tells an agent what to do in every state. The states and possible actions are described declaratively. See, e.g., https://research.ibm.com/projects/ai-planning

One approach to building a planner is to reduce the given planning problem to the satisfiability problem (SAT). This has been done for standard planning solutions, i.e., strong plans (that work against an adversarial environment) and strong-cyclic plans (that work against a fair environment). They do not handle a newly introduced refinement called best-effort plans (that work against environments that are only known not to be adversarial). The goal of this project is to build a planner for best-effort plans by reducing to SAT.

Requirement to be on campus: No

Supervisor: A/Prof. Xiuying Wang

Eligibility:

• DL/ML/Programming courses grades > 80
• Excellent programming skills on Python.  
• Background knowledge in Topology and PyTorch framework will be advantageous.

Project Description:

Glioblastoma is the most common malignant brain tumor with poor survival. The cells within biological networks present high-order relationships that display variations across different types of tumors.

This project aims to analyze the patterns within complex interconnected cellular systems using simplicial complexes. Students expect to apply Simplicial Neural Networks, Simplicial Attention Networks, and Simplicial Attention Neural Networks to the TCGA dataset, which includes Glioblastoma (GBM) and Low-Grade Glioma (LGG), to capture the differences in high-order interactions among cancers, analyze results from these models and visualize the dynamic graphs and high-order interaction patterns.

Useful papers and links: 

1. Simplicial Neural Networks
2. Simplicial Attention Networks
3. Simplicial Attention Neural Networks

Requirement to be on campus: Yes *dependent on government’s health advice.

Supervisor: A/Prof Chang Xu

Eligibility: Programming Skills: Python, basic knowledge of deep learning

Project Description:

In recent advancements, state space models (SSMs), particularly those designed with efficient hardware-aware architectures like Mamba, have demonstrated significant promise in long sequence modeling, notably in language understanding and visual recognition tasks. However, these models often exhibit large sizes and computation requirements. This project aims to investigate the enhancement of the efficiency of SSMs by leveraging a widely used model compression technique - knowledge distillation.

The primary focus of this research will involve an in-depth exploration of state space models and their applications in computer vision tasks, with a specific emphasis on image classification. A comprehensive literature review will be conducted on knowledge distillation methods, with the objective of identifying and understanding their applications in optimizing Mamba models for vision-related tasks.

The core objective is to develop a novel knowledge distillation approach tailored for enhancing the performance of lightweight SSMs. Through this research, we aim to contribute valuable insights into the synergy between state space models and knowledge distillation, ultimately creating more efficient and high-performing SSMs for vision tasks.

Requirement to be on campus: Yes *dependent on government’s health advice.

Supervisor: Prof. Michael Cahill

Eligibility: Experience running applications on databases, and measuring performance properties of systems.

Project Description:

Reproduce results from a recent paper comparing two state-of-the-art open-source database query engines (Excalibur and DuckDB). Use profiling to measure where time is being spent in each, and based on the results, propose changes to improve performance.

Requirement to be on campus: No

Supervisor: Dr. Wei Bao

Eligibility:

• Knowledge in machine learning.
• Knowledge in federated learning.
• Experience in PyTorch.

Project Description:

With the rapid advancement of artificial intelligence, numerous machine learning models have been employed for data analysis and predictions. While the primary focus of traditional machine learning has been on model accuracy, in recent years, there has been a growing emphasis on the fairness of machine learning.

Compared to centralized machine learning, federated learning is a decentralized framework that offers better privacy protection for participants' training data. In federated learning, the server cannot access the local data of federated learning participants, which makes achieving algorithmic fairness more challenging than in centralized machine learning.

In this project, you are expected to devise an attack mechanism that specifically targets the fairness of the aggregation model learned within the federated learning framework. Additionally, an appropriate defense mechanism needs to be designed to safeguard against the proposed fairness attack algorithm. Furthermore, a comparative analysis and discussion should be conducted to assess the effectiveness of the proposed attack and defense mechanisms in comparison to existing attack and defense approaches.

Requirement to be on campus: Yes *dependent on government’s health advice.

Supervisor: Dr. Hazem El-Alfy

Eligibility: Student took a Machine Learning or Artificial Intelligence class

Project Description:

Diabetic retinopathy (DR) is a medical condition induced by diabetes that may result in retinal damage and blood leakage. This sickness may cause a range of symptoms, varying from minor vision impairment and reaching up to total blindness if not treated immediately.

The purpose of this project is to devise an artificial intelligence tool to detect this condition in its early development by analysing retina images. Early detection of such a condition can result in adequate treatment at the right time and prevent severe symptoms, such as blindness.

To this end, the research team participating in this project will survey the recent literature in the area, choose appropriate large image datasets and develop a deep-learning architecture to analyse these images and identify those that indicate potential severe cases.

The results of this research, if promising, will be published in a reputable computer vision conference or journal.

Requirement to be on campus: No

Supervisor: Dr. Hazem El-Alfy

Eligibility: Student took a Machine Learning or Artificial Intelligence class

Project Description:

People who do not make regular stops while driving for long distances run the risk of getting drowsy. According to recent studies, driver fatigue accounts for 20% of fatal road accidents in NSW. Drivers fail to identify that they become tired until they actually are.

This project aims to devise an artificial intelligence tool to detect this condition as early as possible by developing a non-obstructive method that analyses facial images. In particular, the blinking pattern of eyelids can be used for this purpose. Early detection of driver drowsiness and issuing adequate warnings can help reduce the risk of fatalities.

To this end, the research team participating in this project will survey the recent literature in the area, choose appropriate large image datasets and develop a deep-learning architecture to analyse these images and identify how drowsiness can be detected early enough. We anticipate publishing the results of this research, if promising, in a reputable computer vision conference or journal.

Requirement to be on campus: No

Supervisor: Dr. Qiang Tang

Eligibility: WAM>80 and Undergraduate candidates must have already completed at least 96 credit points towards their undergraduate degree at the time of application.

Project Description:

Stellar is one of the leading decentralized payment solutions, at its core, is a new federated consensus protocol.

In this project, we would like to understand how its actual security and/or weakness, and potential analysis and/or fixes.

Requirement to be on campus: No