Abstract: To implement a speech recognizer for a language in conditions of substantial unavailability of related speech training material the first step (1,2) is, based on related speech training material, a multilingual speech recognizer (2) for a plurality of known languages. The recognizer for such given language (5) is then implemented by interpolation (4) starting from the said multilingual recognizer (2). The recognizer (5) generated in this fashion is susceptible of being subsequently refined based on related speech training material acquired online (4) during later use (FIG.

Abstract: A node in a packet-switching data-transmission network has a processor distributing incoming messages or packets from a receiving buffer to transmitting buffers selected according to routing data established by an updating circuit which algebraically combines incremental delays with respective path delays to obtain total delays assigned to message transmission from the processor over respective transmitting buffers and associated outgoing transmission paths. The increment delays assigned to the respective buffers are calculated by a delay estimator utilizing data from the processor including packet-service times and arrival and departure times of the packets in the various buffers, the path delays assigned to routes extending from the respective transmitting buffers to a terminal node being communicated to the updating circuit via the processor from nodes connected downstream of the transmitting buffers.