Knowledge Distillation from Language-Oriented to Emergent Communication for Multi-Agent Remote Control

arXiv:2401.12624v2 Announce Type: replace-cross
Abstract: In this work, we compare emergent communication (EC) built upon multi-agent deep reinforcement learning (MADRL) and language-oriented semantic communication (LSC) empowered by a pre-trained large language model (LLM) using human language. In a multi-agent remote navigation task, with multimodal input data comprising location and channel maps, it is shown that EC incurs high training cost and struggles when using multimodal data, whereas LSC yields high inference computing cost due to the LLM's large size. To …

abstract agent arxiv communication control cs.ai cs.it cs.lg cs.ni data distillation human knowledge language language model large language large language model llm location math.it multi-agent multimodal navigation reinforcement reinforcement learning remote control semantic type work