Pragmatic Speaker

Pragmatic speaker research investigates how agents, both human and artificial, use language effectively in context, considering factors beyond literal meaning to achieve communicative goals. Current research focuses on developing computational models, often incorporating Bayesian approaches and reinforcement learning, to understand and replicate this nuanced aspect of language use, particularly within large language models (LLMs). This work is crucial for improving human-computer interaction, enhancing the trustworthiness and fairness of AI systems, and advancing our understanding of human communication itself.