A CNN plus bidirectional LSTM with attention reaches 89.7 percent accuracy on a new corpus of sarcastic versus real news headlines, beating a CNN baseline by about five points.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2019 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Sarcasm Detection using Hybrid Neural Network
A CNN plus bidirectional LSTM with attention reaches 89.7 percent accuracy on a new corpus of sarcastic versus real news headlines, beating a CNN baseline by about five points.