Electronic Theses and Dissertations

Identifier

1165

Date

2014

Document Type

Dissertation

Degree Name

Doctor of Philosophy

Major

Engineering

Concentration

Computer Engineering

Committee Chair

Mohammed Yeasin

Committee Member

Russell Jerry Deaton

Committee Member

Bashir I Morshed

Committee Member

Xiangen Hu

Abstract

Sense-making is the ability to connect novel information to the familiar schema (or adapt to an unknown environment) and selection of actions. It can be viewed as a continuous effort to bridge the gap between human ability and agent’s task demand in the context of human-agent (system) interaction. A range of the gaps can be considered in the context of human-agent interaction depending on the agent’s role and functions. We focus on cognitive ability-demand gap for the sake of simplicity as it is critical in designing technology solutions that are adaptive, assistive, and can potentially enhance users experience through collaborative sense-making. The main goal of this study is to pursue new avenues in designing assistive solutions for people with varying degrees of disability. Every disability is unique and effective technology solution would require “assistive thinking” and “collaborative sense-making” between the user and the system (agent). The key idea is to develop a suite of techniques to model the ability-demand gap in order to have a deeper insight about human-agent interaction and leverage it in designing assistive technology solutions. The key objectives are to: (a) model the ability-demand gap in terms of cognitive ability using latent response theories, (b) establish the association of model parameters with cognitive resources and cognitive task demands in gap modeling, and (c) use the gap model for collaborative sense-making to design and develop novel assistive technology solutions. The proposed research connects psychometric ability and task difficulty parameters using the latent response model with the human ability and the task demand, respectively. This connection allows to model ability-demand gap using one parameter Item Response Model (IRM). A natural extension for polytomous response using graded response model (GRM) was also used to evaluate performance on complex cognitive tasks. Pilot studies were performed to estimate ability parameter using simple cognitive task (e.g., simple mental multiplication). To further the understanding, two types of abilities (primary and secondary) a dual task scenario (e.g., complex collaborative task) was considered. Discrepancy in secondary task performance was found to be related to the gap. A variation of response latitude around the particular difficulty level (response attitude) was considered as gap value. With a combination of task response attitude, latitude and response time – we propose a 3D response model of gap estimation. To investigate the source of low cognitive performance in terms of mental resource allocation with task ability and demand multiple resource theory with dynamic shift of working memory resources was considered in precision of mental resource computation, which is considered as a direct correlation to mental workload. Thus, the range of score might represent the mental resource level gap. To study deeper analysis of cognitive task performance that reflects cognitive and collaborative ability-demand gap, Maximal Information Coefficient and Maximal Aasymmetry Score between ability and demand vectors are considered. Studies were performed to understand the effect of ability demand gap in collaborative sense making, cognitive dissonance and overload to advance the concept of assistive thinking in designing technology solutions for people with disability. This research expected to have an increased understanding of the parameters to have a deeper insight about collaborative sense making, provide effective feedback, classification of agent’s roles in human-agent interaction and potentially transform assistive technology.

Comments

Data is provided by the student.

Library Comment

Dissertation or thesis originally submitted to the local University of Memphis Electronic Theses & dissertation (ETD) Repository.

Download

Share

COinS