Cooperative Multi-Task Semantic Communication for Joint Classification and Regression Tasks
This paper extends the cooperative multi-task SemCom framework to jointly handle heterogeneous classification and regression tasks on the complex Cityscapes dataset and adopts the information maximization (InfoMax) principle so that it accommodates mixed discrete and continuous semantic variables.
Ahmad Halimi Razlighi, M. Rahman, Maximilian H. V. Tillmann et al.
· 0 citations