Andy Schlaikjer
Faculty Advisor: Luis von Ahn

Title: A Dual-Use Speech CAPTCHA: Aiding Visually Impaired Web Users while Providing Transcriptions of Online Audio Sources

   
     
Short
Bio
 

Andy Schlaikjer is a doctoral student in the Language Technologies Institute at Carnegie Mellon University, where he has been involved in various fields of research, from Question Answering and Information Retrieval, to applied Information Extraction. His primary interests revolve around issues of information overload and accessibility found in many common uses of technology. Before coming to Carnegie Mellon, he received his bachelor's degree from Columbia University in 2002, and was then a member of Columbia's Natural Language Processing group until 2004.

     
Project Synopsis
 

To combat fraud and spam, web services such as online banking, email, and even blogs have adopted security measures which attempt to discriminate between human users and automated software agents. Known as HIPs "Human Interactive Proofs" or CAPTCHAs "Completely Automated Public Turning test to tell Computers and Humans Apart", these technologies are primarily based on visual perception tasks, making them nearly impossible for visually impaired users to perform.

To better address the needs of blind and low-vision users, this project aims to develop an alternative CAPTCHA based on a speech transcription task. In addition, as a side-effect of its use, the CAPTCHA will provide transcriptions of speech found in online audio streams, affording greater indexing and retrieval possibilities for these resources. Once completed, the system will made publicly available online to encourage web services to adopt more accessible CAPTCHA technologies.