Computer science

Supervisor: Dr Vera Chung

Eligibility: WAM>85 and Undergraduate candidates must have already completed at least 96 credit points towards their undergraduate degree at the time of application

Project Description:

Reinforcement Learning (RL) is emerging as a powerful tool in the financial sector, with promising developments in its use for automated trading strategies. By enabling agents to learn from and adapt to dynamic market conditions, RL offers the potential to outperform traditional rule-based approaches.

The project will involve creating and testing RL models in simulated trading environments, focusing on optimising strategies based on real-time market data. Alongside the practical work, a literature review will be conducted to analyse existing RL-driven trading approaches, highlighting their effectiveness, challenges, and applicability to financial markets.

The goal is to evaluate how well RL can be leveraged for trading and identify key areas for future research and development.

Requirement to be on campus: No

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 that serve as recharging stations or delivery destinations 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, which multiple drones may share at the same time to transit between nodes. However, managing drone traffic in congested urban environments presents significant challenges.

The risk of aerodynamic interference among multiple drones operating in shared skyway segments may impact drone delivery efficiency. As a result, it may impact the smooth operation of drones in skyway network.

This project focuses on developing real-time dynamic traffic management algorithms to optimize drone routes and minimize congestion and potential inter-drone interferences. We leverage advanced machine learning-based techniques and real-time data analytics to ensure the proactive detection of potential interference zones. The goal is to dynamically adjust drone routes based on status of the drone traffic in potential interference zones.

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

Supervisor: Prof. Athman Bouguettaya

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

Project Description:

Logistics companies are increasingly adopting drone-based delivery services to meet the growing demand for fast and ubiquitous goods delivery. These drones or Unmanned Aerial Vehicles (UAVs) operate within a skyway network. This skyway network is an interconnected set of nodes and segments. The nodes are the building rooftops which serve both as delivery destination and recharging stations for drones to recharge for long-mile deliveries.

The segments serve as a shared aerial space for drones transit between nodes. However, drones may have different flight speeds, manoeuvrability, and battery capacities as they operate in segments. These differences can lead to inter-drone interference, potentially disrupting the efficient operation of the drones.

This project focuses on developing models and simulations that can replicate the interference between heterogeneous drones in a skyway network. The insights gained from these simulations will inform the development of robust and efficient drone traffic management strategies.

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

Supervisor: Prof. Athman Bouguettaya

Eligibility: WAM>75 and Undergraduate candidates must have already completed at least 96 credit points towards their undergraduate degree at the time of application

Project Description:

Social media has become an integral part of our lives, with users constantly uploading content across various platforms. Unfortunately, some of this content includes untrustworthy images. Traditional approaches to detecting fake images often rely on image processing, which can be costly and computationally demanding. More recent methods that analyse comments on posts may also fall short, as even fake posts can receive supportive comments.

We propose leveraging the metadata of an image, along with the associated post information, to assess its trustworthiness. As part of our approach, we will develop a comprehensive database of online images that can be queried based on their metadata. This database will later serve as a crucial resource for determining the trustworthiness of images. For instance, users will be able to search for images with specific metadata, such as shutter speed, which can aid in identifying potential inconsistencies and modifications in the images.

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

Supervisor: Prof. Athman Bouguettaya

Eligibility: WAM>75 and Undergraduate candidates must have already completed at least 96 credit points towards undergraduate degree at the time of application

Project Description: As social media continues to evolve, the challenge of identifying fake images becomes increasingly complex. Traditional methods that rely on image processing or basic metadata analysis often fall short in capturing the intricate semantics that differentiate genuine images from manipulated ones.

We propose using advanced models such as RoBERTa, T5, and GPT to assess the trustworthiness of online images by learning their underlying semantics. These models will be trained to understand the deeper contextual meanings within images, enabling more accurate detection of fake content. By leveraging this sophisticated semantic analysis, we aim to improve the reliability of identifying untrustworthy images on social media platforms.

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

Supervisor: Prof. Athman Bouguettaya

Eligibility: WAM>75 and Undergraduate candidates must have already completed at least 96 credit points towards their undergraduate degree at the time of application

Project Description:

The proliferation of AI-generated images presents a significant challenge in verifying their authenticity. Traditional image verification methods often fall short against the advanced techniques employed by generative AI, making it difficult to distinguish between genuine and AI-created images. We propose an innovative approach to address this issue by focusing on the subtle traces that generative AI algorithms leave in the metadata of images.

By detecting and analyzing these indicators, we aim to assess the trustworthiness of AI-generated images with greater accuracy. Our method will involve developing specialized tools to identify these metadata patterns, enabling us to differentiate between authentic and AI-generated images more effectively. This approach not only enhances the detection of untrustworthy content but also contributes to maintaining the integrity of online visual media. Through this research, we seek to provide a reliable solution for the growing challenge of AI-generated image verification in an increasingly digital world.

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

Supervisor: Prof. Athman Bouguettaya

Eligibility: WAM>75 and Undergraduate candidates must have already completed at least 96 credit points towards their undergraduate degree at the time of application

Project Description:

Energy Service involves the wireless transfer of energy among IoT devices, harvested from sources like kinetic activity or body heat. Crowdsourcing these energy services can create self-sustained environments, offering convenient and widespread power access.

This project aims to design an end-to-end Service Oriented Architecture (SOA) with three key components: energy provider, energy consumer, and super-provider. Providers advertise services, consumers make requests, and the super-provider manages energy exchanges.

While existing research has addressed consumer-related challenges, this work focuses on the provider's and super-provider's perspectives. It explores strategies to overcome provider resistance to energy requests and applies these services in drone swarm delivery to enhance efficiency.

Additionally, the proposal examines efficient energy provisioning and Quality of Experience (QoE)-based approaches, including utilizing partial services and a recommendation-based system. The proposed QoE-aware framework aims to balance energy distribution and ensure high consumer satisfaction by addressing service availability challenges.

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

Supervisors: Dr Zhanna Sarsenbayeva, Dr Anusha Withana

Eligibility: WAM>75 and Undergraduate candidates must have already completed at least 96 credit points towards their undergraduate degree at the time of application

Project Description:

The aim of this project is to create a sensing mechanism to detect if the user is walking or being encumbered (i.e., carrying a bag) while wearing a mixed reality headset. You will need to develop a system that employs the sensors of the headset together with the patterns detected during the interaction with the headset and use this data to detect the situational impairment (i.e. walking or encumbrance).

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

Supervisors: Dr Zhanna Sarsenbayeva, Dr Brandon Syiem

Eligibility: WAM>75 and Undergraduate candidates must have already completed at least 96 credit points towards their undergraduate degree at the time of application

Project Description:

The aim of this project is to create a non-invasive method of enabling torso-based movement in VR that is predicted based on head and controller orientations and positions.

In this project you will first collect ground truth data by integrating a sensor attached to a user's torso with a VR application. You will then use this data to train an existing neural network model with the aim of predicting torso pose based solely on VR headset and controller pose.

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, consider 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 perform 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 consumers' data or provide unreliable performance. Similarly, an IoT service provider may have concerns regarding their consumers' trustworthiness. Malicious consumers may misuse IoT services by sending malicious software. Therefore, the trustworthiness of service providers and consumers needs to be assessed before new service compositions.

For such an assessment, we need to store historical information about service providers and consumers. It is important to investigate approaches that could help guarantee the integrity of the stored data (e.g., use blockchain-based approaches). This project aims to identify or create real-world datasets for evaluating data integrity-preserving approaches, with an additional focus on developing algorithms to detect tampered trust data.

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

Supervisors: Prof. Seokhee Hong, Dr Amyra Meidiana, Dr Yongcheng Jing

Eligibility: Skills Required: Data Structure and Algorithms and Programming (Java, C++, Python, Javascript)

Project Description:

Technological advances have increased data volumes in the last few years, and now we are experiencing a “data deluge” in which data is produced much faster than it can be understood by humans. These big complex data sets have grown in importance due to factors such as international terrorism, the success of genomics, increasingly complex software systems, and widespread fraud on stock markets.

Visualisation is a powerful tool to compute good geometric representation of abstract data to support analysts to find insights and patterns in big complex data sets.

This project aims to design, implement and evaluate new visualisation algorithms for faithful visualisation of big complex data, to enable humans to find ground truth structure in big complex data sets, such as social networks and biological networks.

These new visualisation methods are in high demand by industry for the next generation visual analytic tools.

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 (Python), problem solving, as well as the ability to work independently are required. You should also have excellent ability to read academic literature (publications) and implement and evaluate algorithms.

Project Description:

For fuzzing, one often needs the ability to decompose given sample corpus and recombine them. This requires the ability to parse given inputs with respect to given grammars. Context free grammars are the most common input specifications available. Context free grammar parsers are typically implemented using one of the many parsing algorithms including GLL, GLR, Earley etc.

This project asks you to implement the GLR algorithm for formal parsing and compare it with the Earley and GLL algorithms (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 (Python), problem solving, as well as the ability to work independently are required. You should also have excellent ability to read academic literature (publications) and implement and evaluate algorithms

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 inference, 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

Supervisors: Dr Jonathan Kummerfeld, Dr Julie Ayre

Eligibility: Experience with Python, JavaScript, and HTML.  Strong results in first- and second-year Computer Science units

Project Description:

When doctors and other clinicians write, they usually provide information that is too complex for the general population. To address this issue, the USyd Health Literacy Lab has developed the Health Literacy Editor, a web app that provides evidence-based, objective, real-time feedback on the complexity of written health information. This helps doctors and other medical staff to revise their writing to be more easily understood. The Editor is currently being rolled out to health staff across NSW.

This project seeks to develop new NLP-based components that provide meaningful and actionable feedback to help end-users simplify health information. For example, to check that:

• assessments of text complexity are accurate
• feedback to users is intuitive and easy to interpret
• feedback to users is easy to act on (e.g. through tailored suggested revisions to the text).

The project will involve extensive programming and learning about a range of NLP technologies.

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

Supervisor: Dr Jonathan Kummerfeld

Eligibility: Experience with Python, JavaScript, and HTML.  Strong results in first- and second-year Computer Science units

Project Description:

AI is being widely deployed today, but in many cases, users are unable to see and change the choices AI models make for them. For example, how can you know whether a ChatGPT summary contains all the correct information you need? You could read the entire document to check, but then you don’t need the summary.

This project will explore AI-resilient systems, where the AI model and user interface are designed to give users the information they need to make more informed choices. A few applications are possible:

• An advisor for the board game Diplomacy
• Helping people find patterns in text data
• Programming aides

Student suggestions for specific applications are also welcome (the project is flexible).

The project will involve extensive programming and learning about a range of NLP technologies.

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

Supervisors: Prof. Kalina Yacef, Dr Bryn Jeffries

Eligibility: 4^th year (Honours) and above, excellent programming skills (Java Script and Python), excellent communication skills, some research experience (e.g. honours) as well as tutoring experience are highly desirable

Project Description:
This project aims to integrate AI, specifically ChatGPT-3, to assist tutors in online programming platforms to generate high quality feedback personalised to the student’s learning needs, style and history as they learn programming. The AI bot will adapt to the current context of the question as well as to the student’s history and level of progress captured in an Open Learner Model (OLM).

Note this project will require collaborating with the student working on the OLM project.

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

Supervisors: Prof. Kalina Yacef, Dr Bryn Jeffries

Eligibility: 4^th year (Honours) and above, excellent programming skills (Java Script and Python), excellent communication skills. Some research experience (e.g. honours) as well as tutoring experience are highly desirable.

Project Description:
The aim of this project is to build an open learner model (OLM) in an online programming platform to provide a snapshot of the student’s learning style, history and progress and enable tailored responses to students’ questions as they learn programming. The aim will be to integrate the OLM in an AI bot supporting tutors answer student queries.  

Note this project will require collaborating with the student working on the AI bot project.

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

Supervisor: Dr Anusha Withana, Dr Zhanna Sarsenbayeva

Eligibility: Excellent programming skills, any experience in VR will be an added advantage

Project Description:

This project explores interaction methods for controlling four hands (two real, two robotic) in a virtual reality (VR) environment using non-fixed mapping. Users can switch control between different sets of hands to complete tasks with varying degrees of freedom and attention levels.

The study aims to understand how users allocate attention and manage multitasking when engaging in complex scenarios, such as simultaneous manipulation and coordination tasks. Through detailed analysis of task performance, cognitive load, and user feedback, the project seeks to identify effective strategies for enhancing VR interaction.

Expected outcomes include insights into user preferences for dynamic control, design recommendations for intuitive VR systems, and contributions to the fields of teleoperation and assistive technology. The findings will inform the development of adaptable control mechanisms, offering potential applications in real-world contexts such as industrial automation and remote operation systems.

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

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 Qiang Tang

Eligibility: WAM>75 and Undergraduate candidates must have already completed at least 96 credit points towards their undergraduate degree at the time of application

Project Description: End to end secure communication and online collaboration is critical for ensuring the privacy of individual users and enterprise when its employees are collaborating on platforms that are usually hosted in public cloud. There are very few products that can offer end to end security, Proton is one of the major service provider that offers end to end security for email, online collaboration and more.

In this project, we will analyze the design and implementation of Proton, to identify whether there are vulnerabilities then propose mitigations; if not, we can try to provide formal security proofs/analysis for those real world products.

Requirement to be on campus: No

Supervisor: Dr Qiang Tang

Eligibility: Math/theory maturity

Project Description:
Secret sharing splits one secret into multiple pieces (say n) that any k of them could come together to reconstruct. Due to distributed nature, when some users misbehave and leaks his share, it is hard to hold them accountable, as any share could be useful.  

Traceable secret sharing is a recent primitive that aims to identify which exact piece contributes to a reconstructed secret, thus moving one stepforward towards accountability. We would explore how to improve existing constructions and generalize to design accountability mechanisms in general distributed system.

Requirement to be on campus: No

Supervisor: Dr Qiang Tang

Eligibility: Familiarity of blockchain would be a plus.

Project Description: Immutable is a decentralised gaming platform, we are going to examine the security of its consensus, and general designs of its web3 gaming system.

Requirement to be on campus: No

Supervisor: Prof. Jinman Kim

Eligibility: WAM>75 and Undergraduate candidates must have already completed at least 96 credit points towards their undergraduate degree at the time of application

Project Description:

The advancement of biomedical imaging has enhanced our ability to diagnose and understand diseases. These biomedical imaging modalities capture different information across various biological scales e.g., macroscopic scale with Computed Tomography (CT), microscopic scale using histopathology and, at microscopic level through spatial transcriptomics. Despite these imaging modalities capturing complementary information, however, the relationship between them remains an area of active research.

The aim of this project is to explore and quantify the correlation and association between the complementation imaging modalities across the biological scales. We will focus on the integration between CT and histopathological data for various cancers using public data. The core technical development will be in multi-modal learning architecture to integrate and align the imaging features across the modalities. The project aims to contribute to more accurate and comprehensive diagnostic methodologies, potentially leading to improved patient outcomes.

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

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: Dr Suranga Seneviratne

Eligibility:

• Must be an Australian Citizen.
• Strong familiarity with AI/ML and Python programming.
• Knowledge of signal  processing will be an added advantage.

Project Description:

Traditional static biometrics, such as face recognition, operate in three settings: (i) user identification, (ii) user verification, and (iii) open-world user recognition. In user identification, the most prominent form involves determining a user's identity based on a biometric sample.

The most practical scenario for user identification operates in an open-set setting, where the model either identifies the user or outputs 'unknown user.' User verification typically involves the user providing identity information (e.g., an ID or RFID tag), and the model compares the biometric sample with stored biometric signatures corresponding to the provided user ID.

Open-world user recognition is a more generalized version of user identification, allowing for user re-identification. Here, given a biometric sample, the model may identify the user, find similar samples with contextual information (e.g., where a similar sample was previously encountered), or indicate that no similar sample exists ('unknown user'). Behavioral biometrics are still emerging in all three of these areas.

This project aims to develop self-supervised learning-based feature extractors for multi-modal biometric data streams for user identification and user verification in open-world settings.

Requirement to be on campus: No.

Supervisor: A/Prof. Nguyen Tran

Eligibility: Machine learning Python coding

Project Description: Next-generation mobile networks require accurate wireless traffic prediction for effective resource management. Traditional methods often rely on centralized training, which is inefficient and does not address the distributed nature of data or inter-spatial dependencies.

To address these challenges, we propose Federated Koopman Learning (FedKooL), an innovative approach leveraging Koopman’s theory to handle non-stationarity and capture inter-spatial dependencies. FedKooL introduces a novel optimization framework to learn the Koopman operator from distributed time series data, utilizing low-rank matrices to model dynamics in a low-dimensional linear subspace.

Furthermore, Rockafellar’s envelope technique is employed to manage low-rank parameters, reducing computational complexity. This approach aims to improve prediction accuracy and resource management in mobile networks.

The effectiveness of FedKooL will be validated through rigorous numerical simulations and performance evaluations, comparing it against several state-of-the-art methods in wireless traffic prediction.

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

Supervisor: A/Prof. Nguyen Tran

Eligibility: Machine learning Python coding

Project Description:

The emergence of the Internet of Things (IoT) has led to a surge in data generation, especially in the form of multivariate time series (MVTS), vital for diverse real-world applications. However, these data streams are often susceptible to disruptions caused by anomalous activities, causing serious problems such as cybersecurity breaches or system failures.

Traditional MVTS anomaly detection (MTAD)methods face challenges in dealing with the heterogeneity and privacy concerns inherent in IoT environments. Addressing this, we introduce a novel federated learning framework for efficient, privacy-preserving MTAD.

This framework harnesses the synergy of Koopman operator theory and Reservoir Computing within a federated optimization process, enabling the collaborative development of an MTAD model across IoT devices without direct data exchange. This model is notably resource-efficient, ideal for implementation on resource-constrained devices, ensuring effective localized detection.

The proposed method will be empirically validated across various benchmarks and use cases, demonstrating the advantages of enhancing IoT robustness and security.

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

Supervisor: A/Prof. Nguyen Tran

Eligibility: Machine learning Python coding

Project Description:

Large language models (LLMs), like the newly released Llama 3.1 with 405 billion parameters, have revolutionized various real-world applications. However, their high-end hardware requirements often limit accessibility for many researchers.

Our project aims to democratize the use of such transformative technology through distributed computing. By leveraging multiple GPUs, we enable efficient inference and fine-tuning of LLMs, making cutting-edge AI research feasible and more inclusive.

This approach not only enhances computational efficiency but also broadens the scope of research possibilities by reducing the barrier to entry for utilizing state-of-the-art LLMs.

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

Supervisor: Dr Anusha Withana

Eligibility:

• Experience in 3D printing and 3D design, skills in programming, and interest in making things
• Experience in Gcode handling would be an added advantage

Project Description:

This project explores the expansive potential of 3D printing by integrating advanced computational design tools and innovative hardware enhancements. The goal is to augment traditional 3D printing, which primarily focuses on creating the physical shape of a design, by incorporating additional attributes such as colour, interactivity, and other physical parameters to fabricate realistic and functional outputs.

We are developing sophisticated computational design tools to bridge the gap between hardware and digital design to achieve these objectives. Additionally, we are innovating hardware approaches to enhance the 3D printing process, enabling the creation of more complex and interactive designs.

This project offers a unique opportunity to push the boundaries of digital fabrication technology, significantly contributing to the evolution of how we physicalize digital models and making substantial advancements in the field of additive manufacturing.

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

Supervisors: Dr Anusha Withana, Dr Zhanna Sarsenbayeva

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 electroencephalography (EEG) signals.

However, existing EEG devices are often bulky and inconvenient to use. In addition, they are often not optimised for long periods of use. We want to develop a personalised wearable EEG device that can cater to real user needs to promote device adherence by 3D printing EEG wearable EEG sensors. You will be working on developing computational methods to automate the design process of sensors.

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

Supervisor: Prof. Irena Koprinska

Eligibility:

• Machine learning skills - completed COMP3308/COMP3608 or COMP4318/5318 with D/HD
• Excellent programming skills
• Excellent communication skills

Project Description:

This project is a collaboration with the Royal Prince Alfred (RPA) Hospital. It will explore hospital admissions, readmissions and length of stay for patients who presented to the emergency department using machine learning technique.

The goal is to provide decision support to clinicians to improve patient care, patient flow and hospital resource management. It will use a large dataset provided by the RPA Green Light Institute.

Requirement to be on campus: No

Supervisor: Dr Xi Wu

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

Project Description:

Node mobility, as one of the most important features of Mobile Ad Hoc Networks (MANETs), may affect the reliability of communication links in the networks, leading to abnormalities and decreasing the quality of service provided by MANETs. The mCWQ calculus (i.e., CWQ calculus with mobility) is recently proposed to capture the feature of node mobility and increase the communication quality of MANETs.

In this project, we aim to implement a reasoning system in proof assistant for the mCWQ calculus to prove its correctness. Our specifications and verifications are based on Hoare Logic.

Requirement to be on campus: No

Supervisor: Dr Sri Aravinda Krishnan Thyagarajan

Eligibility: Strong background in programming, and experience implementing complex systems, knowledge in discrete mathematics, cryptographic libraries are a plus.

Project Description:

Multi-party computation (MPC) is the holygrail of distributed applications and recent cryptographic communication models have pushed the barriers in terms of efficiency and security. Now the parties need to compute and output messages exactly once and are not required to maintain any persistent state through the computation. One of the basic functionalities we can realize is distributed randomness generation, where parties contribute to generating a common random value that is both unpredictable and unbiasable.

Recent proposals for MPC in this stateless models have only been theoretically analysed the performance and it remains to be seen how beneficial these models can be in practice. The project involves benchmarking the practical performance of a recent stateless randomness generation protocol by building an end-to-end system. The potential upside of building such a practical system can help take this model into real-life applications and open new research directions.

Requirement to be on campus: No.

Supervisor: Dr Sri Aravinda Krishnan Thyagarajan

Eligibility: Strong background in programming, blockchain related knowledge in cryptographic libraries are a big plus

Project Description:

Blockchain-based cryptocurrencies offer payment services without involving a central-trusted bank. It is well-known that we can build fair payments between a buyer and a seller using smart contracts. But these solutions are expensive and recent advances in smart contract-less solutions have improved both efficiency and costs for the users.

But these solutions are not widely adopted as they are still in the domain of developers and researchers. The goal of this project is to build full end-to-end system for smart contract-less buyer-seller applications. The buyer and the seller may be using their smart phones and the end-to-end system must be OS-agnostic preferably. The system has to allow communication between the two users and these users must be able to read and write transactions on a blockchain. Such a system can help greatly with adoption of blockchains in the real-world with new efficient and cheap solutions.

Requirement to be on campus: No

Supervisor: A/Prof. Chang Xu

Eligibility:

• Research experience in one of the following areas: deep learning, reinforcement learning, or robotics.
• Proficiency in deep learning programming, e.g., PyTorch

Project Description:

Visual imitation learning aims to enable an agent to learn policies from visual demonstrations. It offers a great chance for robots to directly acquire human behaviors from large-scale human-operating videos available on the Internet.

While imitation learning has been effective in replicating human behavior from low-dimensional demonstrations, it often struggles with visual demonstrations due to their less distinguishable representations compared to low-dimensional proprioceptive features.

The primary challenge of visual imitation learning lies in how to extract valuable features from visual demonstrations. This project seeks to develop advanced methods to enhance representation learning within visual imitation learning, bridging the gap between imitation learning from low-dimensional demonstrations and visual demonstrations.

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

Supervisor: Dr Wei Bao

Eligibility: Knowledge in edge computing and algorithm

Project Description:

With the advances of edge computing, computational tasks can directly be run in the edge devices, facilitating a wide range of smart applications. However, due to the limited computational capacity and energy, the devices may not be able to complete computational-intensive tasks. There are two evident solutions: (1) We offload the tasks to a resource-rich computing node (edge server). (2) We reduce the processing time by sacrificing processing quality (i.e., controllable processing time).

In this project, we aim to study the online bipartite matching problem in this edge computing environment. The tasks arrive at the system, which can be offloaded to edge servers. Each task can be processed by one of the edge servers near them. Each task has multiple levels of controllable processing times: If a task is fully completed, it will obtain a full utility; If a task is completed with a lower processing level, a partial utility will be obtained. Our objective is to maximize the overall accumulated utility. We aim to propose a online bipartite matching algorithm to solve the problem with theoretical analysis on its performance.

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

Supervisor: Dr Wei Bao

Eligibility: Strong mathematics background, experienced in distributed computing

Project Description:

In the era of big data and machine learning, the importance of algorithmic fairness has gained increasing attention, emphasizing the need for equal treatment of different demographic groups in decision-making processes. Distributed computing, particularly federated learning, has emerged as a pivotal approach in leveraging decentralized data while preserving privacy. Federated learning enables multiple clients to collaboratively train a global model without sharing their data.

This paradigm is crucial for privacy-sensitive applications, such as healthcare and finance. Despite its advantages, federated learning is susceptible to fairness issues. Biases present in local datasets can propagate and amplify in the global model, leading to discriminatory outcomes. Current fairness-enhancing techniques often fail to address the unique challenges posed by distributed settings, such as heterogeneous data distributions, limited communication, and adversarial attacks.

The aim of this project is to investigate algorithmic fairness in distributed computing environments and develop robust methodologies to mitigate bias and ensure equitable outcomes across diverse demographic groups. You are expected to create distributed computing algorithms that are both robust and fair, validated through various metrics. Additionally, you are expected to conduct theoretical analysis of the proposed methodologies, including proving their generalisation bounds and performing convergence analysis.

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

Supervisors: Dr Anusha Withana, Dr Hossein Moeinzadeh, Prof. Ken-Tye Yong

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

Project Description:

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 find out. 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: https://doi.org/10.1016/j.jwpe.2023.104349

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

Supervisors: Dr Anusha Withana, Dr Hossein Moeinzadeh, Prof. Ken-Tye Yong

Eligibility: Background in Python programming and Deep learning is required. Having knowledge in causal inference methods are encouraged.

Project Description:
This research project explores the cutting-edge integration of Deep Learning and Causal Inference to address a critical challenge in environmental science: imputing missing water quality parameters. Water quality monitoring often suffers from incomplete data due to sensor malfunctions or gaps in manual sampling.

By leveraging advanced machine learning techniques, you will help accurately predict these missing parameters, ensuring comprehensive and reliable water quality assessments. You'll gain hands-on experience with state-of-the-art deep learning models and causal inference methods, contributing to real-world solutions for environmental sustainability.

This project offers a unique opportunity to work at the intersection of artificial intelligence and environmental science, making a tangible impact on water resource management. It’s an exciting chance for you to develop skills in a high-demand field while contributing to the protection of our planet’s most vital resource: water.

You can find relative information on this project in our recent publication https://doi.org/10.1016/j.jwpe.2023.104349

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

Supervisor: Dr Sasha Rubin

Eligibility: HD in Models of Computation or Algorithm Design

Project Description:
Planning is part of the symbolic/logic approach to 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).

The goal of this project is to extend this to modern (decision-theoretic) solutions such as non-dominated solutions and solutions that minimise the maximum regret.

Requirement to be on campus: No

Supervisor: Dr Hazem El-Alfy

Eligibility: Student took a Machine Learning or AI class and has excellent Python coding skills using the Keras or PyTorch library

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 student 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 Hazem El-Alfy

Eligibility: Student took a Machine Learning or AI class and has excellent Python coding skills using the Keras or PyTorch library

Project Description:

Potholes are a common scene in NSW streets after wet weather. It is estimated that the cost of repairing them reached up to $4 billion dollars in 2022. Compensations paid to the owners of damaged cars and handling liability claims also add up to the bill.

This project aims to devise an artificial intelligence software tool to detect potholes in images. Images can be collected by UAVs or traffic cameras, but that is out of the scope. So far, councils have relied on residents to report potholes in their areas, but this is a slow process which results in more cars getting damaged by the time action is taken.

The student participating in this project will survey the recent literature in the area, choose appropriate large image datasets and develop a deep-learning architecture to detect potholes. If promising, we can publish the results of this research in a reputable computer vision conference or journal.

Requirement to be on campus: No

Supervisor: A/Prof. Xiuying Wang

Eligibility: WAM>80, knowledge and experience in deep learning and image processing

Project Description:

Glioblastoma is the most common malignant brain tumor accounting for 47.7% of all cases with poor survival. Biological networks are intricately interconnected and display high-order relationships that differ among tumor types. However, research on graphs focuses mostly on pairwise interactions, failing to capture the intricate interrelationships within complex systems.

This project uses simplicial complexes to decode the intricate interaction patterns within complex interconnected cellular systems. Students expect to apply three specific simplicial complexes relative algorithms for Graph Classification 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.

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

Supervisor: A/Prof. Xiuying Wang, Dr Bowen Xin

Eligibility: WAM>80, knowledge and experience in deep learning and image processing

Project Description:

Modern multimodal medical images are becoming indispensably important to more accurate diagnosis and treatment, and precision medicine. However, acquisition of imaging modalities is often costly and may be invasive. Advanced medical image synthesis models would be used to generate additional imaging modalities to facilitate clinical workflow.

Typical applications include synthetic computed tomography (CT) for MRI-only dose planning for radiotherapy, which aims to reduce patients’ radiation exposure, improve planning precision, and ultimately achieve precision medicine.

This project will develop deep learning models for generating CT images aiming to improve radiotherapy planning. The student will work on multimodal CT and MRI datasets of head-and-neck cancer. The student will gain cutting edge knowledge in the field and hand-on experience and practice on state-of-the-art deep learning models, image analysis and algorithm development.

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

Supervisors: Dr Michael Cahill, Prof. Alan Fekete, A/Prof. Uwe Roehm

Eligibility:  Needed: interest in database internals, ability to read and debug C/C++; desirable: experience using generative AI including carefully writing prompts

Project Description:

There is widespread enthusiasm about the potential for large language models (LLMs) to revolutionise many industries. One task these tools are particularly tuned for is translating between natural language and programming languages.

This project aims to evaluate whether current LLMs can replace a team of software engineers to build a new DBMS from scratch. To do this, each task performed by software engineers, including writing code, adding features to existing code, fixing bugs, performance tuning and reviewing code changes would all need LLM equivalents.

Requirement to be on campus: No

Supervisors: Dr Michael Cahill, Prof. Alan Fekete, Dr Rahul Gopinath, A/Prof. Uwe Roehm

Eligibility: Needed: good knowledge of SQL; desirable: experience installing software, dealing with configuration and dependency difficulties

Project Description:

Recent research has used generating random data and queries, in order to test the query processing correctness in a dbms platform. Particular challenges are choosing queries which will not trivially fail for violating integrity constraints, and determining what the correct output should be (given the original state of the data, and all subsequent changes). This project will aim to check the latest versions of several platforms, to see whether the tools still find bugs, and if so, what kinds of bugs.

Requirement to be on campus: No

Supervisors: Dr Michael Cahill, Prof. Alan Fekete, Dr Rahul Gopinath, A/Prof. Uwe Roehm

Eligibility: Needed: good knowledge of SQL; desirable: experience installing software, dealing with configuration and dependency difficulties

Project Description:

Recent research has used generating random data and interleaved transactions, in order to test the concurrency control correctness in a dbms platform. A key challenge is efficiently determining whether given output is valid, in view of the isolation level requested.

This project will aim to check the latest versions of several platforms, to see whether the tools still find bugs, and if so, what kinds of bugs.

Requirement to be on campus: No

Supervisors: Prof Alan Fekete, Dr Michael Cahill, A/Prof Uwe Roehm

Eligibility: Needed: understanding of query processing techniques such as indices; skill at setting up software and running measurements on it

Project Description:

Recent research in our group has evaluated the effectiveness of the query plan decisions made by the query optimiser of MongoDB. We considered the performance of the plan chosen by the optimiser, and compared it to the performance of other plans when we force the platform to use them.

This project will apply to same methodology to evaluate the query optimiser of various relational platforms, such as PostgreSQL or MySQL.

Requirement to be on campus: No

Supervisors: A/Prof. 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, Statistics and Social Sciences are advantageous

Project Description:

Since the introduction of the Global Digital Health Index (GHDI) in 2019 and the large-scale adoption by WHO in 2023, the former GDHI has been re-established as the Global Digital Health Monitor (GDHM) with new components in the measurement model to monitor DH progress at the country, regional and global levels.

With data from nearly 170 countries collected based on the indicators prescribed in GDHM measurement model. The aim of this summer project is to synthesise and validate the GDHM measurement model and the seven underlying constructs: namely: (1) Leadership & Governance, (2) Strategy & Investment; (3) Legislation, Policy & Compliance; (4) Workforce; (5) Standards & Interoperability; (6) Infrastructure; (7) Services & Applications. Building on our previous research work, there are two analytical tasks:

1. The first analytical task is to assess the reliability and validity of the GDHM survey instruments using Structural Equation Modelling (SEM)
2. The second analytical task is to evaluate the relevance of GDHM for monitoring several SDGs (including SDG3 – Good Health and Well-being) using set-theoretic-based approaches

Requirement to be on campus: No