January
24, 2019 4:00 PM - 5:00 PM OW 222 |
Dr. Peter Overholser, Dept. of Mathematics, OIT | Artificial Neural Networks and Applications - Artificial neural networks have experienced a dramatic increase in effectiveness and prominence over the past decade. I will give a brief description of an artificial neural network, discuss several basic architectures, and show some implementations. Particular attention will be given to an application known as word2vec. As I am just beginning to explore these topics myself, the treatment will be very accessible. |
February 11, 2019 4:00 PM - 5:00 PM OW 222 |
Prof.
Pat Schaeffer, Dept. of Management, OIT |
Factorial Designed Experiments Using Minitab Statistical Software for Laser Weld Process Robustness - Factorial Experiments are orthogonal, statistically designed experiments most often applied in engineering and science for optimization (maximize or minimize) of a particular output response given multiple input factors. However, maximization or minimization may not always be the experimental objective. It may be more desirable to render an output response relatively insensitive, or robust, to one or more uncontrollable factors. This presentation demonstrates factorial experiments and the use of Minitab Statistical Software to a disc drive motor laser welding application wherein robustness to weld gap is the experimental objective.
|
February 14, 2019 4:00 PM - 5:00 PM OW 222 |
Dr. Rosanna Overholser, Dept. of Mathematics, OIT |
Causation,
not just correlation -
Many of the
questions people have are causal. For example, does eating
chocolate while studying cause higher grades? Does using a
non-standard font on a resume cause fewer job interview
requests? One way to answer such questions is with a
randomized experiment. In this talk I will describe a
formal method of determining when causal questions can be
answered in settings where experiments are not possible.
Several examples will be given and the potential for
application in machine learning problems discussed.
You may have heard the saying “correlation does not imply causation,” meaning that if two quantities change together all we can say is that they are correlated, not that one causes a change in the other. For example, the number of students eating ice cream cones is correlated with the number of shorts worn on campus, but we’d probably be reluctant to conclude eating ice cream causes a student to wear shorts or vice versa. |
February 25, 2019 4:00 PM - 5:00 PM OW 222 |
Dr.
Yuehai Yang, Dept. of Natural Sciences, OIT |
How
"Squishy" is a Roll of Graphese, a.k.a., a Single-Walled
Carbon Nanotube? Despite the tremendous application potentials of graphene and carbon nanotubes (CNTs) proposed by researchers in the last two decades, efficient experimental techniques and methods are still in need for conclusive characterizations of their mechanical properties in order to apply them in high accuracy engineering. In this talk, we will be looking at several mechanical/mathematical models about CNT and its compression in radial direction. Young’s moduli (Eradial) of these straight single-walled carbon nanotubes (SWCNTs) have been measured by using atomic force microscopy (image (a)). The experimental results were consistent with the reported theoretical simulation data. The method developed in this mechanical property study has been applied over the years to measure the mechanical properties of other low-dimension nanostructures, such as nanowires and nanodots. This talk will hopefully help visualize and shed some light on how classical mechanical models work in the “nano-realm”. |
March 14, 2019
4:00 PM - 5:00 PM OW 222 |
Dr. Randall Paul, Dept. of Mathematics, OIT | Mathematical
Perspectives and Projective Geometry In modern times the problem of representing a three dimensional scene in two dimensions (as, say, on a computer screen) is handled by using linear algebra and computers. But people have been drawing convincing pictures of three dimensional scenes for centuries---how did they do it? This talk discusses the mathematical underpinnings of perspective drawing and the resulting branch of mathematics, Projective Geometry. Like the more familiar Euclidean Geometry, Projective Geometry starts from a few, very basic axioms and uses them to prove an elaborate collection of theorems. Unfortunately, the connection between this beautiful, classical field of mathematics and the ideas behind prospective drawing is now not so clear. In this talk we’ll do some perspective drawing, so bring a pencil and a ruler. I also hope to illustrate (as it were) the connection between perspective drawing and Projective Geometry. |