Biomedical Named Entity Recognition at Scale

12 Nov 2020 Veysel Kocaman David Talby

Named entity recognition (NER) is a widely applicable natural language processing task and building block of question answering, topic modeling, information retrieval, etc. In the medical domain, NER plays a crucial role by extracting meaningful chunks from clinical notes and reports, which are then fed to downstream tasks like assertion status detection, entity resolution, relation extraction, and de-identification... (read more)

PDF Abstract

Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT BENCHMARK
Medical Named Entity Recognition AnatEM BLSTM-CNN-Char (SparkNLP) F1 89.13 # 1
Medical Named Entity Recognition BC4CHEMD BLSTM-CNN-Char (SparkNLP) F1 93.72 # 1
Named Entity Recognition BC5CDR Spark NLP F1 89.73 # 2
Medical Named Entity Recognition BC5CDR BLSTM-CNN-Char (SparkNLP) F1 89.73 # 1
Medical Named Entity Recognition BioNLP13-CG BLSTM-CNN-Char (SparkNLP) F1 85.58 # 1
Medical Named Entity Recognition JNLPBA BLSTM-CNN-Char (SparkNLP) F1 81.29 # 1
Named Entity Recognition JNLPBA Spark NLP F1 81.29 # 1
Medical Named Entity Recognition LINNAEUS BLSTM-CNN-Char (SparkNLP) F1 86.26 # 1
Named Entity Recognition LINNAEUS Spark NLP F1 86.26 # 1
Named Entity Recognition NCBI-disease Spark NLP F1 89.13 # 2
Medical Named Entity Recognition NCBI-disease BLSTM-CNN-Char (SparkNLP) F1 89.13 # 1
Medical Named Entity Recognition Species800 BLSTM-CNN-Char (SparkNLP) F1 80.91 # 1

Methods used in the Paper


METHOD TYPE
Residual Connection
Skip Connections
Dense Connections
Feedforward Networks
WordPiece
Subword Segmentation
Layer Normalization
Normalization
Scaled Dot-Product Attention
Attention Mechanisms
Adam
Stochastic Optimization
Linear Warmup With Linear Decay
Learning Rate Schedules
Weight Decay
Regularization
Dropout
Regularization
Attention Dropout
Regularization
Softmax
Output Functions
Multi-Head Attention
Attention Modules
GELU
Activation Functions
BERT
Language Models