Computer Science Departmental Seminar Series

Spring 2007

TCCW 137B

Dr. Ahmed Emam	Feb 6, 2007	Eye Tracking form A to B
Dr. Lakshmi Narasimhan	Feb 15, 2007	Learning, Accreditation and the ABET Processes
Betsy Williams	March 1, 2007	Effectively exploiting HMD technology to explore large virtual environments.
Dr. Zhonghang Xia	March 20, 2007	A Kernel Method for XML Document Classification
Dr. Art Shindhelm	April 3, 2007	A Project Oriented AI Class

Eye Tracking form A to B

Abstract

Our eye movement in general is an unconscious process that makes the ability to visualize a sequence of fixations. Eye-tracking system determines the pupil's orientation, which allows analyst to determine the 3D coordinate of the subject's point of gaze. Eye tracker data used to determine where the subject is looking and eye-gaze data is stored digitally to determine the behavior of the test subject and can be processed in real-time to facilitate interactive feedback. Eye tracking is an primarily filed of human computer interaction science. An introduction to eye tracking system concept, installation, calibrations, strengths of the technology, usability, and some areas of application of eye tracking will be demonstrated.

Learning, Accreditation and the ABET Processes

ABSTRACT

V L Narasimhan

N.B.: “ABET accreditation is NOT about ranking a University or a program ”

-- From http://www.abet.org/

In this seminar I will cover the three topics of learning, accreditation and the ABET processes in accrediting. As many (tertiary) educators would have been well-aware, learning is a continuous process and arises as a consequence and confluence of different styles of education. I will briefly cover the different styles of education at various levels, followed by the need for accrediting the learning and/or educating processes. ABET (Accreditation Board for Engineering and Technology) is the apex body for accrediting undergraduate programs in science, engineering and technology in the USA [1] . ABET is recognized by the Council for Higher Education of all states and their accreditation requirements are rather extensive. In this seminar, I will give an overview of their current process and their proposed new process, which is to become effective from 2008-09.

ABET has now become ABET Inc. in order to have world-wide visibility and impact. However, there are profound differences in the style of learning and the form/nature of education around various countries. The ABET's equivalent body in the EU (European Union), called the Euro-ACE, employs very different process for their accreditation. Their differences stem from their profound divergence in their underlying learning and educational processes. I will touch upon these issues, along with the wider ramifications of the new ABET Inc. to higher education in science, engineering and technology.

As noted, the ABET's accreditation is “ process-driven ” and I shall accordingly draw upon parallels from the ISO 9000 and ISO 9001 Standards that deal with quality issues in the (manufacturing) industry. However, ABET does not employ ISO or equivalent Standards - nor are they envisaging to develop appropriate International Standards on educational processes. This may be the seed for a debate!

Biography of the Speaker

Lakshmi narasimhan obtained his PhD from the University of Queensland , Australia , where he also worked as a Postdoctoral fellow, lecturer and then as a senior lecturer. Lakshmi moved to the Defense Science and Technology Organization (DSTO) as a Principal Research Scientist and then to the University of North Texas , Dallas , as a full Professor in late 2000. He also served as a full professor and Chair in Software Engineering at the University of Newcastle, Australia. Presently he is the Chair of Computer Science at the Western Kentucky University . His current research interests are in: Software Engineering (testing, visualization, instrumentation, product line engineering & CBSE), Information Engineering (medical informatics, mobile systems, information management, information extraction & information fusion) and Computer Architecture (embedded system, reconfigurable computing & performance analysis). Lakshmi is a Senior Member of the IEEE, ACM, Fellow of ACS, IEAust and IEE (UK). He is a Technical Member (representing USA ) of the Expert Panel of ISO (International Standards Organization) and MIMOSA (Machinery Information Management Open Standards Organization, USA ).

[1] The word ABET does not contain the word ‘science', as I believe that it was added relatively recently.

Effectively exploiting HMD technology to explore large virtual environments.

Abstract:

Virtual environments provide people with the opportunity to experience environments remote from their actual physical surrounding. Since head-mounted display (HMD) systems hold the promise of being readily available to the public within the next several years, constraints of the system need to be identified and addressed. In this talk I will discuss how we can understand the perceptual constraints of an HMD virtual environment and leverage that knowledge to build better, more effective environments. In particular, a major drawback of HMD-based systems is the likely limited amount of space available for exploration. My talk will explain how to leverage the natural ability of people to maintain spatial awareness in an HMD-based virtual environment under two conditions: (1) the virtual environment is larger than the physical space containing the HMD system and (2) the method of locomotion used to explore the environment is physical walking.


Brief Biography:

Betsy Williams is a Ph.D. Candidate in Computer Science at Vanderbilt University.  Her research involves developing and evaluating systems of learning in virtual environments. Her broader research interests lie in computer graphics, animation, and image processing.  She regularly serves as a guest lecturer at Vanderbilt University for advanced graphics and animation classes.  She is currently seeking a faculty position for the 2007-2008 school year.

A Kernel Method for XML Document Classification

Abstract

Extensible Markup Language (XML) has been used as a standard format for data representation over the Internet. Unlike the flat text document, the XML document has no vectorial representation, which is required in most existing classification algorithms. In this talk, we propose a kernel method with which a group of XML documents are represented by a set of pairwise similarities. While our similarity measurement successfully preserves structural information of XML data, the derived similarity matrix is not positive semi-definite, which may lead to poor classification performance. We propose a Newton-type method to compute a positive semi-definite matrix nearest to the similarity matrix.