Sparse Rewards Can Self-Train Dialogue Agents.

Barrett Martin Lattimer, Varun Prashant Gangal, Ryan McDonald, Yi Yang

Browse the full ACL paper archive.