External Research Grants

This project aims at developing a software engine that enables planners to generate optimised routing plans based on a given set of data inputs subject to constraints. This work is an offshoot from the Collaborative Urban Delivery Optimisation (CUDO) project which was completed under the Fujitsu-SMU Urban Computing and Engineering (UNiCEN) Corp Lab, that was funded by the National Research Foundation (NRF). The project will customise and extend the functionalities of the CUDO engine developed at SMU, and integrate the engine into Y3 Technologies’ enterprise platform system via an Application Programming Interface (API).

Deep learning has enabled significant advances in applications involving image, text and audio data, in applications such as surveillance, machine translation and speech recognition. These successes have positioned deep learning as a promising approach to address critical challenges (reducing defects, design-time and down-time) in the advanced manufacturing and engineering (AME) domain. A major barrier to this broader application of deep learning is its need for large, labeled datasets to obtain good performance. Thus, the team aims to develop novel deep learning methods that can learn with 10 to 100 times less data as compared to current approaches. This project will enable deep learning to be used in a wider range of applications especially where data is scarce or expensive to obtain. The team will also demonstrate their methods using real-world data from the three identified AME applications (tentatively defect identification, predictive maintenance and circuit design) to show progress and applicability to the AME domain. This project is a collaboration between A*STAR’s I2R, SMU, SUTD, NTU and NUS.

There is an increasing need to use technology to address unequal access to justice, however, this is met with a dearth of empirical evidence to guide the design of access to justice solutions and evaluate their effectiveness at meeting actual needs. This project thus aims to address this problem, through the design and testing of a legal needs questionnaire for Singapore and conducting a pilot survey.

This project investigates 1) when and why skilled mid-career employees in Singapore, and particularly those coming from families of lower socioeconomic status (SES), fail to leverage job opportunities, and 2) how can such workers leverage job opportunities more effectively. In stage 1 of the project, we aim to advance the understanding of job preferences and choices of workers from lower-SES families using a longitudinal study, and to compare them with preferences of local hiring organizations to pinpoint ways in which the two can be more effectively matched. In stage 2 of the project, we will develop a new field intervention that can to help workers effectively seek information and resources to leverage job opportunities, as well as boost their individual coping strategies. Together, this project aims to extend the body of knowledge on workers coming from lower-SES backgrounds, who are at present relatively neglected, and at the same time to have a positive practical impact by improving career outcomes of such workers.

Collaborative machine learning is an appealing paradigm to build high-quality machine learning (ML) models by training on the data from many parties. However, concerns over trust and security have hindered the sharing of data between organizations. This projects aims to address the notion of trustability (i.e. trustable sharing for AI) of the research team’s proposed model-sharing framework in terms of three key principles: assurance of quality and security, fair distribution of benefits, and the right to be forgotten. A fourth aspect, namely the legality of the sharing transactions, is also a key trustability consideration for a learning collaborator partaking in a sharing transaction with others, given that there are many emerging regulatory frameworks to protect data privacy globally. To close the loop between the legal and technical aspects, the team will also study the technical feasibility of current legality frameworks in light of the technological developments of model-centric sharing. The team will design and develop a trustable, operational and legally-compliant model-centric sharing platform called XModelHub based on the above research, and validate with real-world data and applications provided by industry partners. Given that many real-world AI solutions required collaborations between self-interested parties, the team believes that such a platform would be instrumental for powering a data-driven AI innovation ecosystem through trusted sharing for a vibrant digital economy.

Through the project, the team proposes an integrative program of fundamental research towards a vision in which every human will have such an AI assistant for daily life and work. The overall aim is to build conceptual understanding of human-AI collaboration, to develop representations, models, and algorithms for situated assistance, and to integrate them in an experimental device platform for evaluation. The research program consists of five thrusts: (i) situated language communication with reasoning, (ii) visual-linguistic situation understanding, (iii) human collaboration modeling, (iv) robust situated teaming, and (v) the integrative showcase project The “Other Me”, which develops and evaluates the experimental platform Tom – a wearable situated AI agent that assists the human in creating novel artifacts or diagnosing faults. The research aims to empower a new generation of AI assistants for situated, just-in-time, and federated assistance. They will reinvent the relationship between the human and the device in our daily life and work. This research/project is supported by the National Research Foundation, Singapore under its AI Singapore Programme (AISG Award No: AISG2-RP-2020-016). Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not reflect the views of National Research Foundation, Singapore.

As AI becomes ubiquitous, people’s trust in AI is actually dwindling. The key barrier to adopting AI is no longer technical in nature, but more about gaining stakeholders’ trust. Federated Learning (FL), in which training happens where data are stored and only model parameters leave the data silos, can help AI thrive in the privacy-focused regulatory environment. FL involves self-interested data owners collaboratively training machine learning models. In this way, end-users can become co-creators of AI solutions. To enable open collaboration among FL co-creators and enhance adoption of the federated learning paradigm, this project aims to develop the Trustworthy Federated Ubiquitous Learning (TrustFUL) framework, which will enable communities of data owners to self-organize during FL model training based on three notions of trust: 1) trust through transparency, 2) trust through fairness and 3) trust through robustness, without exposing sensitive local data. As a technology showcase, we will translate TrustFUL into an FL-powered AI model crowdsourcing platform to support AI solution co-creation.

In safety-critical applications (e.g., where human lives are at stake), it is crucial for humans to be well-trained to handle expected and unexpected scenarios of varying complexity. Paramedics frequently respond to life-or-death situations, and they must be trained to handle expected and unexpected situations with respect to patient condition and response to treatment effectively. Manned vessels in maritime traffic operate in environments with many other vehicles, and humans must be trained to safely navigate varied situations. A wide range of critical activities in crime response, healthcare, defence, and construction also require training to improve safety. Through this project, the team intends to develop and assess Explainable and tRustworthy (ExpeRt) AI (or Agent) Training Programs (ATPs) with feedback interfaces to adaptively train human(s) for safety-critical applications with showcase projects on emergency response and maritime navigation. The ExpeRt ATPs will generate safe, unexpected scenarios that adapt to observed learner deficiencies and yet provide fair and comprehensive coverage of all cases/situations. This research/project is supported by the National Research Foundation, Singapore under its AI Singapore Programme (AISG Award No: AISG2-RP-2020-017). Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not reflect the views of National Research Foundation, Singapore.

Singapore Management University (SMU) introduced in 2015 a major pedagogical programme of experiential learning for undergraduates, called “SMU-X.” In four years, it has grown from a pilot initiative with a limited number of courses, offered by a small team of enthusiastic and dedicated faculty, to a large programme reaching more than 3,600 student enrolment per year. As SMU-X enters its fifth year, this TRF project aims to evaluate the extent to which the programme has achieved its objectives. The evaluation results will help SMU to (1) enhance its SMU-X initiative for undergraduate students, (2) extend the concept to graduate students, and (3) to share its knowledge and expertise in experiential learning with other institutions of higher learning in Singapore and the region. The research will also contribute to a better understanding of how we should prepare students for a professional environment. Finally, we hope that our research may make a methodological contribution, helping others to measure the impact of experiential and problem-based learning methods.

Drawing on the research of Social Identity Theory and Identity Integration literature, this longitudinal study seeks to investigate the relationship between female students’ Gender-Professional Identity Integration (G–PII) and their creativity in STEM fields. The findings will inform the strategies of structuring STEM programs including the arrangement of mentors and the design of class exercise to incubate female students’ performance.