Speech technologies for under-resourced languages of the world.

Over the past decade, computer speech processing of under-resourced and minority languages has experienced a significant progress. In this paper, we present an analytical review of existing problems and approaches in the field of speech recognition for many spoken languages lacking speech and text resources, including languages of the Russian Federation. The definition and characteristics of under-resourced languages and challenges connected with their automatic processing are presented, as well as projects and investigations dealing with analysis and preservation of under-resourced languages of the world are described.