Attention-based Hierarchical Variational Information Bottleneck for Robust Multi-Agent Communication under Variable Bandwidth
This work proposes AH-VIB, an attention-based autoregressive variational communication model that combines a variational information bottleneck (VIB) with sequential message generation and a hierarchical robustness loss, and evaluates it on a custom cooperative object-inspection and occupancy-mapping task.
Lukas Koch Vindbjerg, Qi Zhang, Yury Brodskiy et al.
· 0 citations