The training and testing splits for our Covid-19 dataset (for entity and event trigger extraction) are released in the ./covid_data/ folder. There are 25 fine-grained entity types and 14 event types (described in Table 6 in the paper.)
The AMR parser used in our model is derived from (Astudillo et al. 2020) in our model. We train the AMR parser on the biomedical corpus on the AMR official website, and the code and the pretrained model are publicly available. Please follow the README file in the code directory to install and use it.
We will release our model code and more detailed running instructions soon. Thanks so much for your patience.