STANCE aligns learned character encodings via optimal transport and scores them with a CNN, outperforming baselines on four of five new alias detection datasets and improving cross-document coreference by 2.8 B^3 F1.
Ventura, Rebecca Nugent, and Erica R.H
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2019 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Optimal Transport-based Alignment of Learned Character Representations for String Similarity
STANCE aligns learned character encodings via optimal transport and scores them with a CNN, outperforming baselines on four of five new alias detection datasets and improving cross-document coreference by 2.8 B^3 F1.