Modeling Non-Cooperative Dialogue: Theoretical and Empirical Insights
This work addresses the challenge of strategic conversations in AI dialogue systems, which is incremental as it builds on existing cooperative models by incorporating non-cooperative dynamics.
The paper tackles the problem of modeling non-cooperative dialogue by developing a theoretical framework and empirical analysis to identify non-cooperative interlocutors during a visual-dialogue task, with results validated through reinforcement learning experiments on a new corpus derived from the GuessWhat?! dataset.
Investigating cooperativity of interlocutors is central in studying pragmatics of dialogue. Models of conversation that only assume cooperative agents fail to explain the dynamics of strategic conversations. Thus, we investigate the ability of agents to identify non-cooperative interlocutors while completing a concurrent visual-dialogue task. Within this novel setting, we study the optimality of communication strategies for achieving this multi-task objective. We use the tools of learning theory to develop a theoretical model for identifying non-cooperative interlocutors and apply this theory to analyze different communication strategies. We also introduce a corpus of non-cooperative conversations about images in the GuessWhat?! dataset proposed by De Vries et al. (2017). We use reinforcement learning to implement multiple communication strategies in this context and find empirical results validate our theory.