International students showcase their minor theses from home
For the first time ever, international students in the Faculty of IT were able to complete the Minor Thesis program from their home countries.
This change was brought into effect because many students have been required to continue their studies from their home countries due to COVID-19 restrictions.
On 29 June 2022, the faculty held the bi-yearly minor thesis showcase. Eight international students dialled in from their home countries to join the other 24 master’s students as they presented their theses as posters.
‘It’s always a pleasure supervising our minor thesis students because they’re some of our best students. They’re expanding the boundaries of knowledge,’ said minor thesis chief examiner Dr Reuben Kirkham.
The students covered a wide array of topics including mental health, music, astrophysics, geopolitics, data protection and more. Many of the projects were closely aligned with the three key challenges identified in Monash University’s Strategic Plan, Impact 2030: geopolitical security, climate change and thriving communities.
Another highlight this year was seeing a large number of students get their papers published, including international student Zheng Fang who has had a paper published in the Q1 journal Entropy and the Q2 journal Econometrics. He has two more under peer review for Q1 journals and has nine citations so far based on Google Scholar.
Zheng’s minor thesis topic is ‘Minimum Message Length in Time Series Applications and Deep Learning’. He proposes a new hybrid time series model selection technique combining deep learning and the information theory of Minimum Message Length, producing better value-prediction forecasting performance with fewer errors.
‘The key to success for a thesis is the research, which is why I found that pursuing this program remotely was not an issue. I communicated with my supervisors once a week via Zoom and it was very efficient,’ said Zheng.
He was supervised by Associate Professor David Dowe, Associate Professor Shelton Peiris (University of Sydney) and Professor Dedi Rosadi (Gadjah Mada University). Zheng hopes to continue researching time series, deep learning and FinTech through a PhD at Monash.
Here are some of the other innovative topics presented on the day:
Automatic Accessibility Documentation in the Built Environment
Student: Ying Zheng
Supervisors: Dr Reuben Kirkham
Ying’s project focuses on automatic accessibility documentation to improve the wellbeing of people living with disabilities. She engaged 10 volunteers and monitored their movements and barriers encountered through a GoPro, feeding the results through a machine learning program to train the AI for better barrier detection.
Her research is valuable because there are currently no fully automated approaches for documenting accessibility in built environments, most relying on direct inspection by inadequate experts or crowdsourcing.
‘AI is the future. I’m very lucky to have had Reuben as my supervisor. My background is in information technology learning, and when I shared with Reuben that I was interested in deepening my knowledge in AI, he interviewed me and took me in,’ said Ying.
Multi-Agent Reinforcement Learning for Electric Vehicles with Network Constraints
Student: Kadawatha Arachchilage Nethmi Vihara Kadawathaarachchi
Supervisors: Dr Hao Wang
Electric vehicles (EVs) offer an environmentally-sustainable alternative to standard combustion vehicles, however the more EVs are used, the higher the energy demand.
Vihara’s project explores a solution using a multi-agent EV charging scheduling algorithm that leverages reinforcement learning, a novel approach to the extant stochastic scheduling algorithms.
‘I could not have asked for a better supervisor. I work full-time and Dr Hao and his co-supervisor PhD student Jiarong were incredibly supportive from the start. We had weekly discussions where they helped me develop my thesis. I’ve learned that there’s always room for improvement and perseverance is key,’ said Vihara.
Detecting the Topic Changes in Streaming Conversational Dialogue Text
Student: Tingting Dong
Supervisors: Dr Greg Rolan, Associate Professor Campbell Wilson
Working with the AiLECS Lab, Tingting’s research focuses on developing a model that automatically generates detected topics in conversations by identifying key words and segmenting them into different categories – a novel approach to traditional AI projects which require manual data labelling. It can be used in text message conversations, audio interviews, automatic indexes and unsupervised annotations.
A key way this model will be used in the future is with AiLECS Lab’s work in combating child exploitation. It can analyse big conversations between children and adults to detect any exploitative material more efficiently.
‘At the beginning it was really tough. I was familiar with supervised learning such as deep learning, so this was my first time investigating unsupervised learning. I was reading four to five research papers a week. But I think overall it was a great challenge,’ said Tingting.
Monash postgraduate students in the Faculty of Information Technology have the opportunity to apply to pursue a research project during their degree through the Minor Thesis program. Learn more about the program on the Faculty of IT website.