In some example embodiments, there may be provided a method that includes receiving, by a user equipment, at least a reporting configuration from a network, wherein the reporting configuration is associated with at least a first group of channel measurement resources to be measured by the user equipment and a second group of channel measurement resources for which beam indexes or beam measurements are predicted by a machine learning model comprised at the user equipment; providing, as an input to the machine learning model comprised at the user equipment, at least a first set of measurements on the first group of channel measurement resources; and reporting, to the network, channel measurement information, wherein the channel measurement information contains at least an index representing a channel measurement resource of the second group of channel measurement resources. Related systems, methods, and articles of manufacture are also disclosed.