Ana Lawry Aguila research guest at EMCL

Ana is a PhD student from the COMBINE Lab at University College London (UCL) visiting the EMCL group as part of a research exchange. Her research focusing on developing unsupervised multimodal autoencoder models for biomedical applications. These methods learn useful low dimensional representations of multiple modalities of input data without requiring labels, which, in the medical field, are often missing or poorly representative of the data at hand. Unsupervised multimodal autoencoders can be applied to the study of heterogeneous neurodegenerative diseases such as Alzheimer’s disease or Epilepsy for personalised medicine, to aid early diagnosis, or identify disease biomarkers. During her time in the EMCL group, she will be working on developing a code library of multimodal autoencoder methods alongside Alejandra Jayme. There exists in the literature, many ways of extending autoencoders to multiple views or modalities. Existing code is often implemented using different Deep Learning frameworks or coding styles. As such, this library aims to collate these existing methods into an easy-to-use library.