APPARATUS AND METHOD FOR LEARNING FROM DATA HAVING MISSION VALUES

There is provided a device (112, 203) for a communication network, the communication network comprising a plurality of input nodes (101, 102, 103, 201, 202), each input node for processing a respective first data (105, 106, 107, 208, 209) relating to an entity (104, 207) and outputting a respective second data, the device for implementing a plurality of neural networks (109, 110, 111, 206a, 206b) and for receiving (701) a respective signal from at least one of the nodes, the respective signal is determined from an evaluation of a quality (116, 117, 211) of one or more of the respective first data and the respective second data; in accordance with the or each respective signal, one of the neural networks is selected (702) to process respective second data received by the device from one or more of the nodes. This may cause the device to infer attributes of the entity when reliable data cannot be obtained from all input nodes.