Listener Model for the PhotoBook Referential Game with CLIPScores as Implicit Reference Chain
This addresses the problem of multimodal communication in collaborative games for AI systems, though it is incremental as it builds on existing methods like DeBERTa and CLIP.
The paper tackled the challenge of predicting shared images in the PhotoBook referential game without requiring reference chains, achieving >77% accuracy on unseen data, which outperforms baselines by over 17 points.
PhotoBook is a collaborative dialogue game where two players receive private, partially-overlapping sets of images and resolve which images they have in common. It presents machines with a great challenge to learn how people build common ground around multimodal context to communicate effectively. Methods developed in the literature, however, cannot be deployed to real gameplay since they only tackle some subtasks of the game, and they require additional reference chains inputs, whose extraction process is imperfect. Therefore, we propose a reference chain-free listener model that directly addresses the game's predictive task, i.e., deciding whether an image is shared with partner. Our DeBERTa-based listener model reads the full dialogue, and utilizes CLIPScore features to assess utterance-image relevance. We achieve >77% accuracy on unseen sets of images/game themes, outperforming baseline by >17 points.