Referential communication in heterogeneous communities of pre-trained visual deep networks
This addresses the challenge of interoperability among autonomous agents with different AI architectures, though it is incremental as a first step in this direction.
The paper tackled the problem of enabling heterogeneous pre-trained visual networks to communicate about objects in their environment, showing that they can self-supervisedly develop a shared protocol for referential communication, which works to some extent for unseen categories and can be learned by new networks with ease.
As large pre-trained image-processing neural networks are being embedded in autonomous agents such as self-driving cars or robots, the question arises of how such systems can communicate with each other about the surrounding world, despite their different architectures and training regimes. As a first step in this direction, we systematically explore the task of referential communication in a community of heterogeneous state-of-the-art pre-trained visual networks, showing that they can develop, in a self-supervised way, a shared protocol to refer to a target object among a set of candidates. This shared protocol can also be used, to some extent, to communicate about previously unseen object categories of different granularity. Moreover, a visual network that was not initially part of an existing community can learn the community's protocol with remarkable ease. Finally, we study, both qualitatively and quantitatively, the properties of the emergent protocol, providing some evidence that it is capturing high-level semantic features of objects.