Our research concentrates on developing new modelling languages, inference methods, and machine learning solutions that support dynamic decision making in complex environments, under uncertainties, and with limited resources.
Our work aims to advance understanding of and solution for dynamic decision making by separating cognitive level with logical and visual languages, from mathematical representation of underlying probabilistic decision models. The cognitive level facilitates human-centered knowledge acquisition and understanding; the mathematical representation supports efficient inference and automated learning from large databases.
We adopt an integrative approach based on Bayesian theory, decision theory, statistical learning, and cognitive theory that serves as a unifying methodology foundation. Our projects are motivated by and tested in a wide range of biomedical and health care settings – from understanding biological processes and systems, to improving patient care and analyzing epidemiological policies and plans in cancer, heart disease, asthma, head injury, and other critical care conditions. Other relevant domains addressed include homeland security, game AI, etc.
Current Research Projects
On-going research in two major projects funded by MOE in Singapore and MIT GAMBIT Lab to support human-centered, adaptive decision making in complex domains. Main challenges addressed include:
- reasoning at multiple level of details and targeted at multiple, distributed users;
- learning with sparse information from multiple, heterogeneous sources;
- adapting to temporal and environmental changes;
- integrating deterministic and uncertain information; and
- combining multi-modal information, e.g., 3-D magnetic resonance images, text-based information, and physiological signals.
Application domains include biomedical and clinical decision support, game AI, robot planning in large and complex terrains, and general commonsense reasoning.
Completed Research Projects
Some major completed projects include:
- Clinical Decision Support Systems, as part of the Medical Informatics Research Program at NUS, funded by the then National Science and Technology Board (now A*Star) and the Ministry of Education (MOE) in Singapore. This work developed a set of multi-perspective, dynamic decision modeling, machine learning, and time-critical decision making techniques in various biomedical domains.
Outcome: Publications; some research results led to a spin-off company from the research group.
- Intelligent Prognostic Analysis in Medicine, funded by the Biomedical Research Council of A*Star and MOE in Singapore. Prognostic analysis emphasizes effective use of information to improve quality, reduce variation, and manage resources in health care procedures. This work developed computational techniques that support cost-effective prognostic analysis using both genomic and phenotypic information, and prototype applications that automate practice guideline and outcome model generation in significant and time-critical health care domains.
Outcome: Publications; follow-on led to translational research on a trial clinical decision support workbench implementation.
- The ResEasy Project, a translational research project funded by the Infocomm Development Authority (IDA) and The Enterprise Challenge (TEC) in Singapore. Project focused on facilitating best practices in process management, guideline execution and outcome analysis in the prospective care of Asthma patients, and acute care of Acute Respiratory Distress Syndrome (ARDS) patients. Collaborators include the National University Hospital, Singapore National Asthma Program, Gleneagles Hospital, and Hewlett Packard. This work developed an open, adaptive trial workbench that supports cost-effective integration, visualization, analysis, security and communication of the relevant information; the workbench can be adapted to different diseases or conditions, and be deployed in multiple sites.
Outcome: Publications; open source toolboxes released in 2008 on Sourceforge. Over 500 world-wide downloads to-date, with continuing recent activities. Trial was one of the first public-private research programs in Singapore, aiming to be a novel implementation model that facilitates and expedites technology transfer. (Go to Reseasy page)