Imagine you are making a conference call on your phone (or computer), with people from different parts of the world. Ideally, each person can speak with their native tongue and each listener will hear the conversation in his/her native language. With the current advances in automated speech recognition (ASR) and machine translation (MT) technology, this is becoming a reality. But there are huge challenges not just from the technological and quality perspective of the speech recognition and translation, but also from the user interface aspect.

The main problem we want to tackle here is designing a good user interface that would enable a smooth Speech to Speech Translation user experience in a one on one call in a desktop PC environment. What would the user experience be if the speech recognition and translation systems are both perfect and always return accurate results? Given the imperfections of the translation and speech recognition systems, how would the UI model change to allow users to continuously provide feedback and to personalize a system?

Team Members

• Daniel Mejia V
• Kenny Risk
• Angapparaj Kalimuthu
• Alex Cheng