I completed my PhD at Cornell University in 2017. Before my PhD, I did a four year integrated masters BSc MMORSE degree at the University of Warwick, and graduated with First Class Honours in 2012.
My research interests are in LSH algorithms, machine learning, statistical theory, and educational pedagogy. I am currently exploring using Monte Carlo methods to complement random projections and other LSH algorithms in conjunction with machine learning algorithms to increase their predictive power without incurring a large speed tradeoff. I am also constantly seeking ways to improve the teaching of courses to undergraduates.
I am currently involved in the following projects.
- “Big Data” and Theoretical Calculations Aided Molecular Design of Fluorophores: from Trial-and-Error to Molecular Engineering (IDG31800104)
This is joint and ongoing work with Liu Xiaogang, Richmond Lee, and Michinao Hashimoto. This work is funded by SUTD-MIT International Design Centre.
- Identifying bottlenecks in teaching and learning mathematics at university
This is joint and ongoing work with Wong Wei Pin, Nachamma Sockalingam, Sergey Kushnarev, and Tan Da Yang in understanding the difficulties that students face when transitioning to university mathematics, in order to develop a mathematics course which caters to their needs. This work is funded in part by the SUTD Faculty Early Career Award.
- Autograder for R Code
I have been involved in developing an autograder for R code since becoming a TA for Statistical Computing in Spring 2016 at Cornell University. The autograder is written in such a way that a TA with minimal programming experience should be able to use the autograder to grade code.
- Keegan Kang, Wong Wei Pin (2018). “Improving Sign Random Projections with Extra Information“. In Proceedings of the 35th International Conference of Machine Learning volume 80 of Proceedings of Machine Learning Research, Stockholm, Sweden, Jul 10-15, 2018, pp. 2484-2492.
- Keegan Kang (2017). “Random Projections with Bayesian Priors“. In , pp. 170-182.
- Keegan Kang (2017). “Using the Multivariate Normal to Improve Random Projections“. In , pp. 397-405.
- Keegan Kang and Giles Hooker (2017). “Control variates as a variance reduction technique for random projections“. In , pp.1-20.
- Keegan Kang and Giles Hooker (2017). “Random projections with control variates“. In , pp. 138-147.
- Keegan Kang and Giles Hooker (2016). “Improving the Recovery of Principal Components with Semi-Deterministic Random Projections“. In , pp. 596-601.
- Keegan Kang and Giles Hooker (2016). “Block Correlated Deterministic Random Projections”. 5th Annual International Conference On Computational Mathematics, Computational Geometry and Statistics.
- Ben Athiwaratkun and Keegan Kang (2015). “Feature Representation in Convolutional Neural Networks“, arXiv preprint arXiv:1507.02313.
- 10.004 Advanced Math II, cohort instructor (Trimester 3, 2018)
- 10.007 Modelling the Systems World, cohort instructor (Trimester 1, 2018)
- 10.004 Advanced Math II retake class (Trimester 3, 2017)
- 10.004 Advanced Math II, cohort instructor (Trimester 3, 2017)
Service and Outreach Activities
Honours and Awards
- SUTD Faculty Early Career Award (2017)
- Outstanding Graduate Teaching Assistant (2017)
- Giving to Warwick student prize (2011)
- Giving to Statistics student prize (2011)
- Warwick Advantage gold award (2011)
- When I was an undergraduate student, I wrote a few revision guides and scribed lecture notes for some mathematics and statistics courses, which can be found here. They may still be of some use.
- I once appeared on the BBC.