Student
Ruchen Liu

Ruchen is working with Novartis on mathematical methods for differential privacy in clinical research.

Ruchen is working with Novartis on mathematical methods for differential privacy in clinical research.

Ruchen graduated from Anhui University with a Bachelor in Economic Statistics in 2019 and completed his MSc in Statistics at the University of Leeds in 2020. His MSc dissertation focused on the comparison of MLE and Bayesian estimator. Besides maths, he enjoys all kinds of sports (including ping-pong, basketball and swimming, badminton and is always open to new sports), cooking, eating delicious food, sightseeing and all other beautiful aspects of life.

Research project title:
Mathematical methods for differential privacy in clinical research

Supervisor(s):
Matthew Nunes and Sandipan Roy

Project description: 

Differential Privacy (DP) is a novel probabilistic framework to provide provable guarantees of data privacy when individual participants submit data for analysis.  Generative (deep learning) models are useful to analyze and generate data but are still prone to adversarial attacks. An open research question is whether generative (deep learning) models can be used in a differentially private framework to either (i) produce private synthetic image and tabular data, which could then be shared safely or (ii) trained in a differentially private manner so that the model itself can be shared. Since our application of interest is clinical trials data, another goal of the project will be to develop DP counterparts of clinical trial methods and quantifying the reliability of analysis results using these methods.