Login

Publications  •  Project Statistics

Glossary  •  Schools  •  Disciplines
People Search: 
   
Title/Abstract Search: 

Dissertation Information for Insuk Oh

NAME:
- Insuk Oh

DEGREE:
- Ph.D.

DISCIPLINE:
- Library and Information Science

SCHOOL:
- Rutgers University (USA) (2006)

ADVISORS:
- Mark Frank

COMMITTEE MEMBERS:
- Nicholas Belkin
- Daniel O'Connor
- Matthew Stone

MPACT Status: Fully Complete

Title: Modeling believable human-computer interaction with an embodied conversational agent: Face-to-face communication of uncertainty

Abstract: Embodied conversation agents (ECA) are computer-generated humanoid substitutes that are used to facilitate human-computer interaction. In the course of these interactions, ECA's must often express uncertainty to the human user. Thus, endowing an embodied conversational agent (ECA) with the ability to express and recognize uncertainty becomes a key challenge. The primary goal of the research was to develop an integrated framework for studying human behaviors for ECAs to achieve believable interactions, starting with the expression of different levels of uncertainty. The research consists of four distinct stages: Stage I, to elicit spontaneous uncertainty expressions; Stage II, to examine the reliability of these signals within and across communication channels; Stage III, to analyze these expressions to identify their components (part 1) and then code these behaviors into an ECA (part 2); and Stage IV, to examine whether the coded behavioral signals sent by the ECA are still recognized by human users. The results suggest there are reliable behavioral clues for degrees of uncertainty, and that these clues seem to be present in the nonverbal Communicating Uncertainty with an Embodied Conversational Agent behavior of people, and that they can be coded into ECA's such that humans interfacing with the ECA perceived the ECA's expression at the same level of uncertainty as was coded into it. Taken together, this research provides not only empirical evidence of a reliable uncertainty expression that can be coded into ECA's, but suggests a framework for coding other human cognitive and emotional states into ECAs to facilitate the believability of interaction between a human and an ECA.

MPACT Scores for Insuk Oh

A = 0
C = 0
A+C = 0
T = 0
G = 0
W = 0
TD = 0
TA = 0
calculated 2008-07-07 15:25:58

Advisors and Advisees Graph

generating graph, please reload

Students under Insuk Oh

ADVISEES:
- None

COMMITTEESHIPS:
- None