Spoken Language Systems
We research the design, implementation and evaluation of Spoken Language Systems. These systems allow users to interact with computers using speech. One of our research initiatives is focused on Distributed Listening (DL). DL uses multiple machines in a distributed architecture to perform speech recognition and spoken language understanding. We also do research on multimodal user interfaces. These environments use multiple modalities to interact with users.
We are also investigating how to speech-enable databases, allowing users to interact with databases using speech. An aim of this research is to perform Information Verbalization. Information Verbalization is the use of computer supported, auditory interactions to amplify understanding of abstract and/or large data. In general, we are conducting research on natural interactions that use spoken language and multiple modalities to perform human-computer interaction.
Advanced Learning Technologies
We are building adaptive learning environments that allow educators to create instruction and provide learners with personalized instruction. These systems interact with people using multimodal interfaces, some use Animated Pedagogical Agents (APA).
We are building game-like interfaces that provide naturally interactive instruction using animation, artificial intelligence, and speech.
Usability
We are investigated holistic evaluation methods for computing systems. Currently, we are working on a holistic evaluation method called the holistic usability measure (HUM) for voice user interfaces (VUI).
Databases and Data Analytics
We use data mining to investigate business intelligence, education and societal issues. We also have projects that investigate data warehousing and information management.
Culturally Relevant Computing / Ethnocomputing
Ethnocomputing deals with the identification of the embeddedness of culture within computing artifacts and the process of applying culture to these artifacts by design to influence use. We specialize in the design, implementation and evaluation of culturally relevant computing systems.
Our team works in the new Research Institute for the Study of Diversity at Auburn University.
AADMLSS:
AADMLSS City Stroll is an interactive game-like environment that uses culturally relevant cues, gestures, sounds and lyrics to teach students algebra. AADMLSS City Stroll consists of three individual components; Instruction, Practice and Assessment. Each component has a specific purpose that leads to the next component. The Instruction component engages the student in interactive, game-like lessons on algebra concepts situated in familiar surroundings. The Practice component reinforces the Instruction component by providing one on one tutoring with an animated tutor using speech. The naturally interactive tutoring is modeled after human to human tutoring. The Assessment component assesses the student's comprehension of the algebra concepts using basic testing strategies that would appear on a standardized test. Combining gaming technologies with artificial intelligence is providing a unique multimedia learning experience for students to learn Algebr!
a. Currently, AADMLSS City Stroll is under going a pilot study at an inner city high school in the Midwest.
Ernest V. Cross
Priyanka Gupta
Yolanda McMillian
Philicity Williams
BOTA Think Tank:
The 2006 Think Tank will explore the social, economic, and political status of Blacks in the 21st century. This call for proposal seeks proposals and subsequent papers that address solutions for health, economic, and educational disparities as they affect the African American community. The Think Tank intends to mobilize the collective contributions and genius of scholars, activists, and community organizers interested in moving beyond rhetoric to providing Black communities with ideas, plans, and programs that result in revitalization efforts. This year's theme, “Black Scholars and the Study of Black Folks: Setting Interdisciplinary Research and Policy Agendas for the 21st Century,” will address the myriad of issues associated with improving the status of African Americans in the United States.
Ernest V. Cross (lead)
Cary Woods Project
In addressing students with learning deficiencies it is often useful to approach the students in a manner that does not directly emphasize the students’ weaknesses, but instead focuses on their strengths. The SimBuilder Science Project targets fourth and fifth grade students with reading deficiencies, with the goal of improving reading comprehension through science, by creating computer based simulations using the SimBuilder Squeak e-learning tool. The SimBuilder Squeak software allows novice programmers to create educational simulations from scratch. The students were given a science challenge based on their grade level, and asked to create a simulation that would present a solution to this challenge. Students were also given weekly readings that were directly related to their science problem, and a journal in which they responded to “thinking like a scientist” questions.
Alexandria Williams (lead)
Kenneth Rouse
Kinnis Gosha
Cross-cultural predictive usability metrics:
The prevalence and frequency of World Wide Web using have been more than before and ever growing. Because of the multicultural and international nature of the Internet, designing interfaces suitable for cultural diverse users becomes an important research problem in HCI. In the dissertation research work, a Cross Culture Predictive Usability Metrics will be developed. The main focuses are (1) creating an effective Cross-Cultural model to map the famous Hofstede’s five culture dimensions into User Interface Design; (2) developing a predictive usability metrics generation system based on the Cross-Cultural model. Additionally, the mechanics of the predictive usability metrics generation system will not only produce metrics neutral to every, or a majority of, ethnic group(s), but also can be fine tuned to favorite specific culture(s). The resulting system should be an affordable and easy-to-use Web system to guide interface designers on Inter-Cultural !
design and provide usability ratings for their mockup systems.
Yifang Gu (lead)
CSSE Document Management System:
Computer Science and Software Engineering Document Management System
(CSSEDMS) is a web application designed for all CSSE faculties and
students. It acts as a central depository for all CSSE digital reports,
theses and dissertations. Through CSSEDMS, CSSE faculties and students
can upload their documents, search and browse documents in the
repository. In general CSSEDMS is a typical web application. Guided by
usability evaluation and modern software architecture ideas, this
project shows the practices to design and implement such a typical web
application to meet user requirements and satisfaction.
Chao Wang (lead)
Data Mining:
Medication Administration Incident Data - Discover the key areas within the medication administration process that contribute to the majority of medication administration errors.
Chao Wang (lead)
Distributed Listening
This study is done using Distributed Listening (DL) and the PRA for an
intra-vehicle speech recognition experiment. The setup for the experiment uses an Acura and 3 tablet PCs to compare the speech recognition experiments against the speech recognition system in the Acura. Noise is created in the car that simulates natural
noise when the car is driving. The results of the study will show if the DL outperforms the in-vehicle speech recognition engine.
Ernest V. Cross
Priyanka Gupta
Yolanda McMillian
David Thornton
Idongesit Mkpong-Ruffin
Holistic Usability Measure
The HUM will provide the evaluators with a complete methodology
to capture and evaluate their speech-enabled systems. It is divided into
various metrics. These metrics may include user opinion, recognition
performance, task completion time, or any other user specified metric.
The designer would be able to place emphasis on any specific metric
that, according to him/her, is most important for the domain of the
specified Spoken Language System. This research will involve identifying
the metrics that impact Spoken Language Systems and methods to quantify
the qualitative metrics. The HUM will also be able to evaluate several
systems which are designed for the same purpose to identify the most
usable one.
Priyanka Gupta
Christin Hamilton
iTech: An Interactive Technical Assistant
Nearly everyone who owns an appliance or device has had the need for technical assistance. This experience usually entails timely searching through a large paper manual or trial and error attempts to come up with an appropriate query to be applied to an online manual. There are two main mediums in which technical communications are currently provided. The first medium is the paper manual. The issues with this medium have been widely documented, especially by technicians in the armed forces. The problems include lack of portability, inaccuracy, and increasing content and complexity (Ventura 1988). As the popularity of the Web grew, online manuals were introduced.
While online assistance reduces the geographical distance between the user and technical documentation, and the issues with inaccuracy and portability; the problems with understandability and specifically search time prevail. The electronic medium also led to a shift towards increased user satisfaction. The requirements for efficient user assistance now include unobtrusiveness; context-sensitivity; consistency and preciseness (Price 1981). Trends towards more user-centered development and a collaborative approach to the field were occurring and resulting in the emergences of hypermedia and interactive applications. However, among these applications, a need for a natural interactive solution still exists.
iTech is an interactive technical assistant that not only addresses the problems with search time, but also provides a conversational interface. Using a new methodology for conversational question answering, called the Answers First Approach (A1 Approach), iTech uses an interactive avatar to address the aforementioned problems.
Dale-Marie Wilson (lead)
Prime III:
The prevalence and frequency of World Wide Web using have been more than before and ever growing. Because of the multicultural and international nature of the Internet, designing interfaces suitable for cultural diverse users becomes an important research problem in HCI. In the dissertation research work, a Cross Culture Predictive Usability Metrics will be developed. The main focuses are (1) creating an effective Cross-Cultural model to map the famous Hofstede’s five culture dimensions into User Interface Design; (2) developing a predictive usability metrics generation system based on the Cross-Cultural model. Additionally, the mechanics of the predictive usability metrics generation system will not only produce metrics neutral to every, or a majority of, ethnic group(s), but also can be fine tuned to favorite specific culture(s). The resulting system should be an affordable and easy-to-use Web system to guide interface designers on Inter-Cultural !
design and provide usability ratings for their mockup systems.
Vince Cross (lead)
Idongesit Mkpong-Ruffin
Yolanda McMillian
Philicity Williams
Kinnis Gosha
QueManager Project:
With the popularity of databases in today’s technology savvy world it is surprising that many aspects of its possible use in everyday life are still unexplored. One example of this is the friendly, neighborhood barbershop. Although problems differ from one establishment to another, a similar problem always occurs in each, “Who is next in line?” To provide a solution to this problem BarberManager was created.
BarberManager is a database driven, online application to manage the waiting list of a barbershop or a beauty salon. This application solves the primary issue in question, such as knowing who the next customer is, and who is scheduled next for a particular barber and much more. Other benefits of using this database oriented technology are that it is advantageous for the customer as well as the business. In order to provide better functionality and to maximize use of all the information collected, BarberManager has three interfaces: the Administrative interface, the Employee interface and the Customer interface . This makes it a robust solution to an existing practical problem.
Kinnis Gosha (lead)
