Compares BERT, ESP, and Cui2Vec embeddings within ESIM on the MedNLI shared-task dataset to assess performance and internal representations for medical inference.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.IR 1years
2019 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
UW-BHI at MEDIQA 2019: An Analysis of Representation Methods for Medical Natural Language Inference
Compares BERT, ESP, and Cui2Vec embeddings within ESIM on the MedNLI shared-task dataset to assess performance and internal representations for medical inference.