Your goal is to develop novel approaches to neural reasoning starting from textually defined tasks in 50% of his/her time and to work on multimodal sarcasm detection in the other 50% time. You will be working with Prof. Alex Binder, Prof. Soujanya Poria, and Prof. Dorien Herremans.

The first task on neural reasoning will include work packages such as:

• to create a novel dataset for neural reasoning with annotations, in parts with help of crowd annotation facilities.
• to implement and validate neural reasoning baselines on CLEVR and on the new dataset.
• To develop novel approaches to neural reasoning.

The second task will also require, aside from dataset construction,:

• the implementation of uni-modal baselines for sarcasm detection
• the development of novel multi-modal for sarcasm detection

The candidate is expected to demonstrate practical knowledge of tensorflow 2.x or pytorch and of approaches in NLP. Both research directions have ample publishing opportunities in high ranked venues for eager and able individuals.

Qualifications, Skills, Experience:

The candidate is required to to have a PhD in computer science or math or physics. Furthermore the candidate is expected to have a track record of publications in well-ranked conferences or known and reputable journals. The position requires an understanding of machine learning and of deep learning beyond having taken coursera courses. This will touch fields such as question answering, graph CNNs, NLP, possibly also RL. The candidate is expected to demonstrate coding skills in deep learning toolboxes (e.g. his/her github). The candidate should be able to work in unix-based environments, and have skills in python, and latex.

Experience in one of the above mentioned research fields is a plus.

To apply, please send your materials to dorien_herremans at domain name [sutd.edu.sg] with subject [postdoc neural]

Given covid-19 restrictions, we can currently only hire candidates that are already in Singapore.