Two-way function computation

Seiyun Shin, Changho Suh

Abstract: We explore the role of feedback for the problem of reliable computation over two-way multicast networks. Specifically we consider a scenario in which there are forward-message computation demands and feedback is offered through the backward network for aiding the forward-message computation. We characterize the feedback computation capacity of a four-node Avestimehr-Diggavi-Tse deterministic network in which two nodes in one side wish to compute modulo-2 sums of two independent Bernoulli sources generated from the other two nodes. As a consequence of this result, we show that the backward network can be more efficiently used for feedback, rather than if it were used for independent backward-message computation. Our achievability proof builds upon a network decomposition framework developed in our earlier work.