IBM has a team of almost 100 employees working on the development of an automatic language translation tool that would be quick, as well as highly accurate to the extent that it would facilitate instant messages between the speakers of two distinct languages. The project is nFluent and at present it is being developed with the intention of teaching computer terminology related to the businesses of IBM. The programmers working on the nFluent project are trying to get a computer to look for and identify statistical patterns that exist between sets of translated content, as well as those which exist among the words.
The company has encouraged its 400,000 work force across over 170 countries to contribute to nFluent. It began by presenting its employees with a worldwide translation challenge in which each employee’s contribution would be measured with the help of a points-based system and the points earned would later be converted to charitable donations. The process involves employees correcting translations produced by nFluent. The challenge mobilized around 6,000 employees to suggest improvements in over 2 million words worth of data for 11 languages that was translated by nFluent. According to an IBM researcher, the parallel data produced by the employees is used to update the models. Until now nFluent has only been used by the employees of IBM, but looking into the future, the company intends to create a finished product that can be sold to other organizations.