Active Learning of Multiple Source Multiple Destination Topologies