TTI-COIN at BioCreative VIII Track 1

Abstract

We built two neural-network based methods that use external data for NER and RE. For NER, We aimed to learn using multiple existing datasets. We propose Conditional VAE (CVAE) with conditions to create slightly different span representations for each dataset. For RE, we constructed a model that integrates the representations of the entities acquired from the neighborhood knowledge graphs, which are subgraphs around the entities, and the representations of the input document.

This article is part of the Proceedings of the BioCreative VIII Challenge and Workshop: Curation and Evaluation in the era of Generative Models.