Practical Applications of Language Technology

With a growing number of practical applications being developed, language technology has become an important branch of the software industry. Such applications include:

  • Machine translation systems: 
    • There are a wide range of commercial systems that automatically translate one natural language into another available today. However, the quality of the translations produced by these systems is still limited, and machine translation remains one of the most important research areas of computational linguistics.
  • Natural-language access systems:
    • These are programs that provide natural language interfaces for databases or expert systems instead of using formal query languages that are hard to learn.
  • Speech recognition and speech synthesis systems: 
    • Speech-driven operation of devices, dictating machines, text-to-speech systems, telephonic dialogue and information systems.
  • Systems for accessing and managing texts: 
    • Document retrieval (e.g. for extensive technical documentation, legal statutes), text filtering (sorting messages from electronic news services customized electronic newspapers), text understanding (partial comprehension of the content of texts and answering questions about it), automatic abstraction (compiling summaries of texts with respect to a specific profile or a specific question).
  • Systems for generating written texts: 
    • Generating technical documentation from a formal description of the respective devices, as well as generating stock market reports, weather reports, etc. from raw data.
  • Intelligent text critique systems: 
    • Grammar and style correction programs.
  • Intelligent computer-aided language learning systems:
    • Common contemporary systems simply compare the student's input with fixed sequences of words and thus judge them as "correct" or "wrong". These systems can be improved by determining the grammatical structure of a sentence entered by a student, evaluating its correctness and locating potential errors as precisely as possible.