Compositional Understanding in Signaling Games
This addresses a fundamental limitation in signaling game theory for researchers in AI and game theory, though it appears incremental as it builds on existing models.
The paper tackles the problem of receivers in signaling games failing to learn compositional information, even when messages are compositional, by constructing models where genuine compositional understanding evolves. It presents two new models, a minimalist and a generalist receiver, that are simpler than previous alternatives and enable learning from atomic message components.
Receivers in standard signaling game models struggle with learning compositional information. Even when the signalers send compositional messages, the receivers do not interpret them compositionally. When information from one message component is lost or forgotten, the information from other components is also erased. In this paper I construct signaling game models in which genuine compositional understanding evolves. I present two new models: a minimalist receiver who only learns from the atomic messages of a signal, and a generalist receiver who learns from all of the available information. These models are in many ways simpler than previous alternatives, and allow the receivers to learn from the atomic components of messages.