I received my Bachelor degree in Biology in Beijing Normal University and my Master degree in Bioinformatics in the University of Copenhagen. My Bachelor thesis aimed at building a database-linked website for post-translational modification analysis. During the Master program I mainly focused on implementing Computer Science or Data Science methods on biological and medical data, which involves programming techniques as well as various concepts and algorithms, such as statistics, machine learning and deep learning.
Machine learning has always been my interest and focus of study. I’ve taken Kaggle competitions for image classification using deep learning and an advanced course on Medical Image Analysis. My Master thesis involved developing, tuning and evaluating a deep-learning model for calling genomic copy number variations by image recognition, which consists of a Convolutional Neural Network and a Bi-LSTM Recurrent Neural Network.
Secondment 1: Employ/extend non-negative matrix (tri-)factorization methods for pattern discovery
Secondment 2 : Learn how to translate knowledge discovery into a user-friendly integrative pipeline
Hunting for patient subtypes through image-based phenotypes as biomarkers for major gene effects in medical disorders.
For my project, I am seeking for a novel pipeline that can integrate multiple data types in order to cluster or subtype individuals with the presence of confounding effects. Because both data modalities I am using, namely facial images and genomic data, are of high dimensions, it is crucial to effectively reduce the dimensionality in the pipeline and to find better data representations as units of analysis. The outcome may bridge the existing market of healthcare and the emerging market of precision medicine in the sense that every individual will be more accurately assigned to relevant strata for risk prediction, diagnostics or disease management.
- Machine learning and Deep learning