Negation detection as a Multi-task learning (MTL) layer

I have been going through this paper - [Joint Entity Extraction and Assertion Detection for Clinical Text](https://arxiv.org/pdf/1812.05270.pdf) - which proposes an MTL approach to negation detection that leverages overlapping representation across sub-tasks i.e., jointly model named entity and negation
in an end-to-end system.

I already have NER model in place and was thinking how would I implement MTL using HMTL but I find it difficult mold given examples in this repo into negation multitask.

@VictorSanh : Would like to know your take on how to go about it?