This project explores new approaches to investigating the fundamental research problem of learning from small (labelled) data, called one-shot learning or few-shot learning, and will develop new algorithms and techniques to devise one-shot learning machines with human-like learning capabilities.
This project tackles the problem of developing intelligent multi-agent planning and decision making algorithms, which can scale to a much larger number of agents and support significantly more complex agent behaviour, than currently possible. The proposed work will propose new models and algorithms that are applicable to a wide range of problems of practical importance, particularly in urban system optimisation.
This project aims to use game theoretic models to randomise patrols and visits conducted by the police, with the goal to minimise their predictability and enhance security efforts.
This multi-disciplinary project brings together the expertise of the School of Social Sciences, School of Information Systems and Lunch Actually (a dating and relationship advice provider), to develop the first holistic AI mobile platform that provides dating and relationship solutions and advice based on comprehensive personality-match models.
The project is a collaboration between the School of Information Systems and Ospicon Systems, a pioneer of the world’s first patented optical fiber based breath-sensing technology for infants and the elderly. By leveraging on the School's and Ospicon Systems’ respective strengths in the areas of IoT data analytics and fiber-optic sensing, the project team seeks to further enhance the performance of Ospicon’s product offerings.
The project seeks to identify misinformation on the web by studying the content and propagation patterns of known cases of misinformation. Through the analysis of these misinformation, the research team seeks to propose methods that will allow for automatic flagging of suspicious pieces of information circulating on the web which can be sent to experts for verification.
The project is a joint collaboration with Acronis Asia Research and Development Pte Ltd, and it aims to develop a non-intrusive monitor that leverages artificial intelligence to achieve protection against data-infringing malware on Android devices. A dynamic analysis framework would be built to perform dynamic analysis of Android apps on un-rooted Android phones to achieve the feature of non-intrusiveness, while a deep learning solution would be developed to identify the specific Android app responsible for performing sensitive operations.
Awarded with a second year funding, the project looks into building fundamental data fusion techniques to combine data from both physical urban sensors and social media sensing to generate improved insights into the evolution of urban events, and a software library of tools that extract and combine analysis techniques across multiple socio-physical sensing channels.
The project, led by Trampolene Limited, aims to design a scalable, self-sustaining system that can collect, classify and determine accessible point-to-point routes that are suitable for barrier-free access.
The project aims to understand MRT events and commuting experience through sentiment analysis of public tweets related to MRT events generated by Singapore's Twitter users.