Generating Diverse and Accurate Visual Captions by Comparative Adversarial Learning

Connections are not prepared for this record yet, or its source metadata has changed. The original record remains available while background snapshots are rebuilt.