About Me
I am a PhD student at the Australian National University, under the supervision of Richard Nock, Ke Sun, and Lexing Xie. I am also a student trainee in Masashi Sugiyama’s Imperfect Information Learning Team at RIKEN AIP.
My research focus is on utilising boosting algorithms, information geometric tools, and the theory of loss functions with a focus on fairness and privacy in machine learning. In this research, I have co-created the pyBregMan open-source Python package with Frank Nielsen. Recently, I have been exploring topics involving generalization bounds and theory involving classification with rejection and importance weighting. Previously I have worked on topics including formal methods / theorem provers, visualisation in academic influence, knowledge graphs, universal approximation theorems, and point process models.
Publications
-
Rejection via Learning Density Ratios
-
Tempered Calculus for ML: Application to Hyperbolic Model Embedding
-
Tradeoffs of Diagonal Fisher Information Matrix Estimators
-
Online Learning in Betting Markets: Profit versus Prediction
-
3D NLTE Lithium abundances for late-type stars in GALAH DR3
-
Fair Densities via Boosting the Sufficient Statistics of Exponential Families
-
Fair Wrapping for Black-box Predictions
-
Interval-censored Hawkes processes
-
On the Variance of the Fisher Information for Deep Learning
-
UNIPoint: Universally Approximating Point Processes Intensities
-
Influence flowers of academic entities