Machine translation involves software to translate text or speech from one language to another. One of the first machine translation companies was SYSTRAN, created by Dr Peter Toma in 1968. 

The European Commission and the U.S. Department of Defense have hired SYSTRAN for significant projects. Up until May 30, 2012, SYSTRAN provided the technology for Yahoo! Babel Fish and other services. Up until 2007, Google's language tools made use of it. The Dashboard Translation widget in macOS makes use of SYSTRAN.

Machine translation

Basic machine translation (M.T.) involves the mechanical replacement of words from one language with words from another. However, this rarely results in a successful translation since recognition of whole phrases and their closest analogues in the target language is required. Moreover, many words have more than one meaning, and not all terms in one language have equivalents in another. In addition, Better translations, addressing variances in language typology, translating idioms, and isolating anomalies are all being achieved using corpus statistical and neural methodologies to solve this challenge.

Modern machine translation software frequently permits customization by domain or profession. This method works exceptionally well in fields that employ formal or formulaic language. Therefore, machine translation of official and legal texts more quickly yields output we can use than machine translation of speech or less regular material. Furthermore, human intervention can also lead to higher-quality work; for instance, specific systems translate more appropriately if the user can identify which words in the text are proper names. These methods have made M.T. an effective tool for assisting human translators, and in some rare instances, it has even produced output that we may use directly (e.g., weather reports).

History

Throughout its history, there has been significant discussion about the development and promise of machine translation. Since the 1950s, several academics, most notably Yehoshua Bar-Hillel, have questioned whether fully automatic machine translation of excellent quality is even possible. In addition, Linux, Solaris, and Microsoft Windows (including Windows Mobile) all support commercial versions of SYSTRAN. Historically, rule-based machine translation (RbMT) was a feature of SYSTRAN systems. SYSTRAN implemented the first hybrid rule-based/statistical machine translation (SMT) technology in the industry with the introduction of SYSTRAN Server 7 in 2010. Furthermore, the company employed 59 people as of 2008, including 15 computational linguists and 26 computational professionals. From 70 employees in 2006 to 59 in 2008, the workforce shrunk.

Conclusion

SYSTRAN, which had its roots in the Georgetown machine translation project, was one of the few machine translation systems to withstand the significant funding reduction following the ALPAC Report in the middle of the 1960s. During the Cold War, the firm was founded in La Jolla, California, to handle text translation from Russian to English for the U.S. Air Force. Under the supervision of the USAF Foreign Technology Division (later the National Air and Space Intelligence Center) at Wright-Patterson Air Force Base, Ohio, a sizable quantity of Russian scientific and technical materials were translated using SYSTRAN. Although only rough, the translation quality was typically sufficient for understanding the subject. Furthermore, the global language market entered a new era during the dot-com boom.