Development of interpretable AI methods for computational biology
A post-doctoral research position is available at the IBM Research – Zurich Laboratory to develop mathematical and computational methods to enhance interpretability in AI methods for computational biology. The successful candidate will work on demonstrating how interpretable artificial intelligence can lead to the generation of new biological insights in computational biology, a flourishing field, where the recent application of deep neural networks to long-standing problems such as the prediction of functional DNA sequences, the inference of protein–protein interactions or the detection of cancer cells in histopathology images has brought a breakthroughs in performance and prediction power. However, despite initial achievements, many of these models were built as black-boxes and have failed to provide new biological insights. A major focus of the work will be to develop new methods that can shed light on the underlying biological principles driving model decisions.
A major application area will be the analysis of single-cell mass cytometry (CyTOF) data from tumor samples. CyTOF technology is achieving tremendously high throughput regarding the number of profiled cells, with some of the most recent datasets consisting of millions of single cells (see, for instance, recent work by our group. The successful candidate will develop new state-of-the-art deep learning networks to model CyTOF data and novel interpretable approaches to extract insight from the models. Several directions will be explored, some of them relying on the generation of disentangled representations that clearly and disjointly map latent and data generative factors; the exploration of architecture-constraint models, which can enhance the interpretability of certain components of the network; or the investigation of parallels with Lagrangian mechanics in physics to identify network invariances that can highlight underlying mechanisms.
The successful candidate will work in a highly interdisciplinary team and be supported by the rich AI community assembled at IBM Research – Zurich. Candidates should have a strong background in Computer Science, Mathematics or Physics and be interested in cancer-related research. Strong programming skills are necessary. Experience in mathematical modelling, statistics, probability and machine learning and/or deep learning are necessary. Candidates with a Mathematics or Physics background are especially encouraged to apply.
About the group
The Systems Biology group at IBM aims to develop new mathematical and computational approaches to analyze and exploit the latest generation of biomedical data. In the context of cancer, the group focuses on developing computational and statistical approaches to unravel cancer molecular mechanisms using high-throughput multi-omics datasets and single-cell molecular data. A major interest is the development of artificial intelligence approaches for personalized medicine and drug modelling, activities that have been supported by two H2020 consortia: PrECISE (2015–2018) and iPC (2019–2022).
Recently, the group has become interested in the field of interpretable artificial intelligence. Although this line of research is relatively young, the group has some preliminary methods that will be presented at NeurIPS 2019 (MonoNet). A particular approach for interpretability, attention mechanisms, has also been used to improve the performance and transparency of drug models (PaccMann). Also, in the summer of 2019, the group organized two workshops on “Interpretability for deep learning in computational biology” at ISMB and BC2 conferences, both of which attracted significant attention within the computational biology community.
IBM is committed to diversity at the workplace. With us you will find an open, multicultural environment. Excellent flexible working arrangements enable both women and men to strike the desired balance between their professional development and their personal lives.
How to apply
Interested candidates please send an application consisting of a CV, list of publications and at least two letters of reference to: Dr. María Rodríguez Martínez