Tingting's paper accepted in AAAI 2023, the premium conference on AI

The paper "DAMA: Learning Delay-Aware Communication for Multi-Agent Reinforcement Learning", co-authored by Tingting Yuan, Hwei-Ming Chung and Xiaoming Fu has been accepted by the Thirty-Seventh AAAI Conference on Artificial Intelligence (AAAI 2023), to be held in Washington DC, USA during Feb 7-14, 2023. With 1721 papers (out of 8777 submissions) accepted after a rigorous double-blind peer-review process, the conference has an acceptance ratio of 19.6%. Congratulations! 

Abstract: Communication is supposed to improve multi-agent collaboration and overall performance in cooperative Multi-agent reinforcement learning (MARL). However, such improvements are prevalently limited in practice due to the existing communication schemes often ignoring communication overheads (e.g., communication delays). In this paper, we demonstrate that ignoring communication delay has detrimental effects on collaborations, especially in delay-sensitive tasks (e.g., autonomous driving). To mitigate this impact, we design a delay-aware multi-agent communication model (DAMA) to adapt communication to delays. Specifically, a TimeNet in DAMA is responsible for adjusting the waiting time of an agent to receive messages from other agents such that the uncertainty of delay can be reduced. Our experiments reveal that DAMA has a non-negligible performance improvement over other mechanisms by making a better trade-off between the benefits of communication and the costs of waiting for messages.