Doctoral Thesis Proposal
Disambiguation of Imprecise User Input Through Intelligent Assistive Communication

Slides

• Google Drive (Preferred)
• PowerPoint
• PDF
• Plain Text

Thesis

Intelligent interfaces can mitigate the need for linguistically and motorically precise user input to enhance the ease and efficiency of assistive communication.

Abstract

Many people with severe speech impairments use icon-based augmentative and alternative communication (AAC) systems. These systems typically present users with hierarchical arrays of icons that are sequentially selected to construct messages, which are then spoken aloud by text-to-speech (TTS) engines. Although ease and speed of message construction are essential, current systems are often slow and require repetitive physical movements that are fatiguing. This dissertation challenges three main assumptions common to icon-based AAC systems. These assumptions influence interface design decisions and place increasing demands on the user rather than the system. The current work leverages natural language processing, machine learning, and context-sensing capabilities to design intelligent communication interfaces that shift the cognitive and physical burden from the user to the system to allow for faster, less fatiguing communication. This work also has broader impact for continuous modalities, such as brain wave and eye gaze activity, for other communication and entertainment applications.

Links to the Documents

• PDF Version of Full Proposal
• PDF Version of Annotated Bibliography (Literature Review)

Thesis Committee

Rupal Patel, Ph.D.
Dr. Rupal Patel is an Associate Professor at Northeastern University and holds joint appointments in the Department of Speech-Language Pathology and Audiology (SLPA) and the College of Computer and Information Science (CCIS). She specializes in understanding neuromotor control of speech disorders and designing and developing assistive communication technologies that leverage the user's capabilities. Given her domain knowledge and experience with the end user group, Dr. Patel will provide overall guidance across several domains including the design of the predictive algorithms and the user interface and the usability studies.

• Role: Thesis Advisor
• Email Address: r.patel at neu.edu

Javed Aslam, Ph.D.
Dr. Javed Aslam is a Professor in the College of Computer and Information Science at Northeastern University. He specializes in information retrieval and machine learning, both areas that are highly relevant to this work. In particular, much of the latest research in language modeling, text prediction, and semantic distance comes from the field of information retrieval, while machine learning is an integral part of predictive behavior and corpus linguistics.

• Role: Committee Member
• Email Address: jaa at ccs.neu.edu

David Smith, Ph.D.
Dr. David Smith is an Assistant Professor in the College of Computer and Information Science at Northeastern University. He specializes in natural language processing and computational linguistics and has worked extensively in the areas of machine translation, information retrieval, and digital humanities.

• Role: Committee Member
• Email Address: dasmith at ccs.neu.edu

Shaun Kane, Ph.D.
Dr. Shaun Kane is an Assistant Professor in the Department of Information Systems at the University of Maryland, Baltimore County. He specializes in assistive technology and human-computer interaction, especially for mobile devices in distracting environments, and has published research in the areas of gestural interfaces and AAC. Additionally, Dr. Kane served on the committee that reviewed an early draft of this work at the Doctoral Consortium of ACM ASSETS 2012.

• Role: External Member
• Email Address: skane at umbc.edu