• DOI:10.18653/v1/N19-1265
Corpus ID: 80628347

Beyond task success: A closer look at jointly learning to see, ask, and GuessWhat

@inproceedings{Shekhar2019BeyondTS,
  title={Beyond task success: A closer look at jointly learning to see, ask, and GuessWhat},
  author={Ravi Shekhar and Aashish Venkatesh and Tim Baumg{\"a}rtner and Elia Bruni and Barbara Plank and R. Bernardi and R. Fern{\'a}ndez},
  booktitle={NAACL-HLT},
  year={2019}
}

Ravi Shekhar, Aashish Venkatesh, +4 authors R. Fernández

Published in NAACL-HLT 2019

Computer Science, Psychology

We propose a grounded dialogue state encoder which addresses a foundational issue on how to integrate visual grounding with dialogue system components. As a test-bed, we focus on the GuessWhat?! game, a two-player game where the goal is to identify an object in a complex visual scene by asking a sequence of yes/no questions. Our visually-grounded encoder leverages synergies between guessing and asking questions, as it is trained jointly using multi-task learning. We further enrich our model via… CONTINUE READING

View on ACL

iris.unitn.it

Save to Library

Create Alert

Cite

Launch Research Feed