Research Interests
Our research activities may be broadly divided into two areas: biodata science and education technology.

Biodata science is an exciting new area where the associative technologies associated with data science and relevant thinking skills are applied onto biological and health-related problems.

These may include how to resolve batch effect problems when effecting large-scale data mergers, improving study reproducibility, and understanding how choice of normalization method impacts downstream data modeling. We are also very interested in understanding how heterogeneity and data holes impacts outcome of analysis, especially in how it affects our interpretation of the underlying biological system.

Our interest in Education technology (EdTech) pertains to the use of data analytics for analyzing student performance and also the development of software for facilitating learning. In particular, we are interested in how we may leverage on big data and machine learning to unravel indicators of human-based deep learning. Instead of fielding our work on typical classroom-type settings, our interest is in combining EdTech with high-impact practices in teaching and learning, where deep learning is more likely to take place. Currently, we focus our research on NTU’s unique Deeper Experiential Engagement Project (DEEP), a large-scale pilot experiential learning project spread across different colleges.

Examples of some recent projects include:

BioData Science and Computational Biology

1. Dealing with confounders in omics analysis
2. Enabling more sophisticated proteomic profile analysis
3. Resolving the missing protein problem using meaningful context
4. Understanding the cost of batch effects in biological big data analysis
5. Developing graph literacy skills
6. How to improve upon weak validation practices in current machine learning

Education Technology
1. Not feeling it — How does sentiment and motivation affect academic performance?
2. Using machine-based deep learning to uncover the signs of human-based deep learning
3. High-impact pedagogical practices