On decoding equations with partial channel state information

We have shown that over interfering links, it is possible to decode a function of the involved messages much more efficiently than the full messages themselves. This is useful for example as a physical layer for network coding. Typically, some degree of channel state information at the receiver is required. In this note, we study the effect of only partial channel state information. For several models of interference, novel achievable rates are derived. These may prove to be useful as a physical layer for non-coherent network codes.