Multi-lingual agents through multi-headed neural networks
This addresses the issue of language incompatibility and forgetting in multi-agent systems, which is incremental as it applies continual learning techniques to a specific domain.
The paper tackles the problem of catastrophic forgetting in multi-agent reinforcement learning where agents forget previous languages when adapting to new ones, and proposes using multi-headed neural networks to enable agents to maintain multiple languages, showing empirical validation in a referential MNIST-based game where the method outperforms existing approaches.
This paper considers cooperative Multi-Agent Reinforcement Learning, focusing on emergent communication in settings where multiple pairs of independent learners interact at varying frequencies. In this context, multiple distinct and incompatible languages can emerge. When an agent encounters a speaker of an alternative language, there is a requirement for a period of adaptation before they can efficiently converse. This adaptation results in the emergence of a new language and the forgetting of the previous language. In principle, this is an example of the Catastrophic Forgetting problem which can be mitigated by enabling the agents to learn and maintain multiple languages. We take inspiration from the Continual Learning literature and equip our agents with multi-headed neural networks which enable our agents to be multi-lingual. Our method is empirically validated within a referential MNIST based communication game and is shown to be able to maintain multiple languages where existing approaches cannot.