Emergent Quantized Communication

The field of emergent communication aims to understand the characteristics as it emerges from artificial agents solving tasks that require information exchange. Communication with discrete messages is considered a desired characteristic, for scientific and applied reasons. However, training multi-agent system not straightforward, requiring either reinforcement learning algorithms or relaxing discreteness requirement via continuous approximation such Gumbel-softmax. Both these solutions result in poor performance compared fully communication. In this work, we propose an alternative approach achieve -- quantization communicated message. Using message allows us train model end-to-end, achieving superior multiple setups. Moreover, natural framework runs gamut Thus, sets ground broader view deep era.