Tamil IT ! : Interactive Speech Translation in Tamil

The Tamil IT! (Interactive Translation) speech translation system is intended to allow unsophisticated users to communicate across the Tamil ↔ English language barrier, without strong domain restrictions, despite the error prone nature of current speech and translation technologies. Achieving this ambitious goal depends in large part on allowing the users to interactively correct recognition and translation errors. We briefly present the Multi Engine Machine Translation (MEMT) architecture, describing how it is well suited for such an application. We then describe our incorporation of interactive error correction throughout the system design. We are currently in the process of developing a Tamil ↔ English system based on this architecture. A Brief Overview of Tamil Language Analysis Like any other language analysis process, Tamil language analysis also involves morphological analysis, syntax analysis and semantic analysis. Tamil is a Morphologically rich language. Most of the grammatical functions are embedded into the word in the form of inflections. Morphological Analysis Here is an example of Tamil morphological analysis. eeRineen eeRu in een (climbed) (Verb) (Past tense) (I Person+Singular+Neuter) In the above example the word eeRineen (climbed) has three morphemes viz.(1) eeRu [Verb for climb], (2) in [Past tense marker] and (3) een [GNP marker]