January Midterm Goals

Since this past summer, I have been working and researching to investigate the problems in the computer recognition of spoken contact names as well as try and pose my own approach to a solution. Today, general speech recognition for normal words, phrases, and sentences is a developed technology especially now with the integration of Artificial Intelligence to learn grammar rules and sentence structure. However, seeing as they do not obey the rules of sentence structure or grammar, names still exist today as a sort of gray area in the capabilities of speech recognition. My main goal for my January Day presentation is to successfully get the point that recognizing spoken names is a problem that needs to be solved across. Along with this, I also want to conduct research and develop my own working model to potentially solve the problem using a new approach of my own. In order to get my audience and the judges to believe this is a problem, I plan on relating it to normal everyday lives, perhaps even theirs. I can talk about how, today, when drivers use the hands-free calling system in their cars they get frustrated and distracted from driving. On top of this, I will mention how, in the future, it is likely that self-driving cars will rely on a speech-based interface for interaction with their users. With this in mind, these cars will need to be able to recognize street names, the names of buildings/structures, and potentially contact names if one's address becomes a part of contact information. I have been working on my own model for some time now, so to give my audience an understanding of it, I will showcase my code and also give a basic crash course on phonetics as the building blocks of language.