DUTIR at the BioCreative V.5.BeCalm Tasks: A BLSTM-CRF Approach for Biomedical Entity Recognition in Patents

Patents contain the significant amount of information. Biomedical text mining has received much attention in patents recently, especially in the medicinal chemistry domain. The BioCreative V.5.BeCalm tasks focus on biomedical entities recognition in patents. This paper describes our method used to create our submissions to the Chemical Entity Mention recognition (CEMP) and Gene and Protein Related Object recognition (GPRO) subtasks. In our method, a bidirectional Long Short-Term Memory with a conditional random field layer (BLSTM-CRF) is employed to recognize biomedical entities from patents. Our best CEMP submission achieves an F-score of 90.42% and our best GPRO submission with type 1 achieves an F-score of 79.19%.