Actor-critic (AC) algorithms have been widely used in decentralized multi-agent systems to
learn the optimal joint control policy. However, existing decentralized AC algorithms either …

Multi-agent policy gradient methods in centralized training with decentralized execution
recently witnessed many progresses. During centralized training, multi-agent credit …

In the last decade, there have been significant advances in multi-agent reinforcement
learning (MARL) but there are still numerous challenges, such as high sample complexity …

While autonomous vehicles (AVs) may perform remarkably well in generic real-life cases,
their irrational action in some unforeseen cases leads to critical safety concerns. This paper …

This dissertation investigates two popular machine learning frameworks, namely, minimax
optimization and multiagent reinforcement learning (MARL). There are a large number of …

Reward shaping has proven to be a powerful tool to improve an agent's performance in
single agent reinforcement learning. Recently, this method has also been applied in multi …

Extracting decentralised policies from joint action-values is an attractive way to exploit
centralised learning. It is possible to apply monotonic value factorisation to guarantee …