About me

I am a PhD student in computer science studying machine learning at the University of Cambridge, supervised by Ferenc Huszár and Bernhard Schölkopf. I am grateful to Cambridge-Tübingen Fellowship and Premium Research Studentships for funding my PhD studies.

My research interests are on scientific induction (a.k.a causality) and non-i.i.d. data. I am interested in building theoretical foundations of causality in non-i.i.d. data: from Causal de Finetti (NeurIPS 2022) to Do Finetti (Under review), from out-of-distribution to out-of-variable (ICLR 2023).

The quest to understand why machines learn relies on the fundamental assumption that train and test data come from i.i.d. data. BUT animals including humans do not perform data shuffling in our life experiences of absorbing information flows. We learn in non-linear, complex, dependent ways. And the quest to understand learning in non-i.i.d. setting has just begun: exchangeable and out-of-variable data offers a realistic and doable next step.

I did Master in Machine Learning with Ricardo Silva at UCL and Master in Mathematics at the University of Cambridge. I was supervised by Richard Samworth during my undergraduate time working on variable selection in high-dimensional statistical inference. I obtained BA in Mathematics with Distinction from the University of Cambridge.