SentiVAE

Code and data associated with the paper

Hoyle, A., Wolf-Sonkin L., Wallach, H., Cotterell, R., and Augenstein, A. (2019). Combining Sentiment Lexica with a Multi-View Variational Autoencoder. NAACL-HLT.

The main script, vae.py, should reproduce the combined representation from the paper.

Since we have not scrutinized the licences for each resource, for the time being we will refrain from reproducing the individual sentiment lexica here. They may be found at the following locations (see paper for full citations):

• SentiWordNet (Baccianella et al., 2010)
• MPQA (Wilson et al., 2005)
• SenticNet 5 (Cambria et al., 2014)
• Hu-Liu (Hu and Liu, 2004)
• General Inquirer (Stone et al., 1962)
• VADER (Hutto and Gilbert, 2014)