Bilingual Sentiment Embeddings accepted to ACL 2018

Bilingual Sentiment Embeddings: Joint Projection of Sentiment Across Languages

I’m really excited to announce that the ong paper that I wrote in conjunction with Roman Klinger and Sabine Schulte im Walde was accepted at ACL 2018. We establish a joint embedding projection and sentiment training method which performs well with as few as 1000 pairs of word translations.

The approach allows us to easily perform sentiment analysis in under-resourced languages and effectively alleviates the data bottleneck.