SOVIET PATENT BULLETIN PROCESSING: A PARTICULAR APPLICATION OF MACHINE TRANSLATION

Abstract

Soviet patents are used for, among other things, gleaning information on developing trends in Soviet technology. Originally  the patents had to be translated by human translators,  indexed manually, and then entered into the CIRC II database for data manipulation and subsequent searching.  This paper presents a description of the task, and of some of the technical problems that had to be overcome in changing to a machine translation system.