Infosys wins ‘Similar Language Translation Task’ Challenge in the 2020 WMT International Machine Translation Competition

Bengaluru, India – September 1, 2020: Infosys (NYSE: INFY) today announced that it has won first place in the ‘Similar Language Translation Task’ category challenge organized by the 2020 World Machine Translation (WMT20). WMT is one of the leading machine translation competitions globally, and a premier platform for researchers to showcase their solutions. Demonstrating advancements in machine translation (MT) technology, Infosys created an Artificial Intelligence (AI) model that helped translate Hindi to Marathi with extreme accuracy, winning the top position in the challenge.

Infosys was evaluated on the performance of the state-of-the-art MT systems for translating between pairs of languages from the same language family. For translation, Infosys came up with a unique deep learning approach that uses novel techniques to achieve a high degree of accuracy even with small data sets. Infosys achieved first place in machine translation for Hindi to Marathi demonstrating the effectiveness of the AI model, given that both are non-English languages, with a high probability of inaccuracies.

This year was WMT’s fifth conference on machine translation and 14th consecutive iteration of the WMT shared task. This competition is organized by researchers from Apple, Facebook, Rochester Institute of Technology, University of Edinburgh, University of Tokyo, Charles University in Prague and other premier institutions.

For further information, please contact: