Generalizing Emergent Communication
This addresses the problem of enabling emergent communication for transfer learning and generalization in AI agents, though it is incremental by building on existing platforms.
The study tested whether deep reinforcement learning can foster a grounded discrete communication protocol between agents in a grid world, finding that proper environmental incentives suffice without specialized techniques and that longer communication intervals encourage more abstract semantics.
We converted the recently developed BabyAI grid world platform to a sender/receiver setup in order to test the hypothesis that established deep reinforcement learning techniques are sufficient to incentivize the emergence of a grounded discrete communication protocol between generalized agents. This is in contrast to previous experiments that employed straight-through estimation or specialized inductive biases. Our results show that these can indeed be avoided, by instead providing proper environmental incentives. Moreover, they show that a longer interval between communications incentivized more abstract semantics. In some cases, the communicating agents adapted to new environments more quickly than a monolithic agent, showcasing the potential of emergent communication for transfer learning and generalization in general.