Abstract: Systems and methods for improving the efficiency of propagating electronic messages are provided. In some embodiments, a feed of activities on a network such as an online social network (OSN) is analyzed to determine influenced activity probabilities and intrinsic activity probabilities for participants in the network. A Helmhotz Green's Function matrix is determined in order to calculate an overall influence weight for each participant in the network. A Woodbury-Sherman-Morrison formula may be used to accelerate computation of the Helmholtz Green's Function matrix, thus allowing efficient updates to the overall influence weights based on newly monitored activities. The participants with the highest overall influence weights may be selected to propagate a new message, thus providing the greatest likely distribution of the message with the fewest originating transmissions.

Abstract: Systems and methods for improving the efficiency of propagating electronic messages are provided. In some embodiments, a feed of activities on a network such as an online social network (OSN) is analyzed to determine influenced activity probabilities and intrinsic activity probabilities for participants in the network. A Helmhotz Green's Function matrix is determined in order to calculate an overall influence weight for each participant in the network. A Woodbury-Sherman-Morrison formula may be used to accelerate computation of the Helmholtz Green's Function matrix, thus allowing efficient updates to the overall influence weights based on newly monitored activities. The participants with the highest overall influence weights may be selected to propagate a new message, thus providing the greatest likely distribution of the message with the fewest originating transmissions.