Our paper “AITL: Adversarial Inductive Transfer Learning with input and output space adaptation for pharmacogenomics” got accepted at ISMB 2020 and will appear at the Bioinformatics journal!

We proposed a method that adapts the gene expression of cell lines and patients and measures of the drug response between them in an end-to-end fashion. AITL utilizes adversarial learning to adapt the input space and multi-task learning to adapt the output space. The motivation behind AITL is that cell lines and patients are very different because of basic biology and different approaches to assess drug response in them. AITL is the first method that can address both input and output spaces discrepancies.

Please read the draft here: https://www.biorxiv.org/content/10.1101/2020.01.24.918953v1

The code, trained model, and the employed datasets are available here: https://github.com/hosseinshn/AITL

ISMB 2020 was supposed to be in Montreal this year, but due to the spread of COVID-19, it will be a virtual conference.