AI RESEARCH
Emergent Communication between Heterogeneous Visual Agents through Decentralized Learning
arXiv CS.AI
•
ArXi:2605.11695v1 Announce Type: cross Symbols are shared, but perception is private. We study emergent communication between heterogeneous visual agents through decentralized learning, asking what visual information can become shareable when agents have different visual representations. Instead of optimizing messages through a shared external communicative objective, our agents exchange only discrete token sequences and update their own models using local perceptual evidence.