## Spring 2003

[ 5/1 | 4/30 | 4/29 | 4/24 | 3/25 | 1/27 | 1/23 | 1/21 ]

**Thurs. 5/1 (12:00) - The Influence of Learning on Chance, History and Adaptation in Artificial Evolution
**Ash Dean

Department of Mathematics and Computer Science

Dickinson College

(Honor's Thesis Defense)

How much of "Survival of the fittest" is actually "Survival of the lucky?" Come to this talk for some insight! Mr. Dean will explain techniques for simulating learning and evolution using computers. He will explain how he used these techniques to investigate how learning impacts the effects of random chance, genetic history and adaptation on the course of evolution.

**Wed. 4/30 (12:30) -Dynamic Correlation: the effect of learning on evolution when learning and evolutionary tasks are different
**Adam Labadorf

Department of Mathematics and Computer Science

Dickinson College

(Honor's Thesis Defense)

How might your ability to learn to compute a partial derivative affect the course of evolution? Come to this talk and find out! Mr. Labadorf will introduce techniques for simulating both learning and evolution on a computer. Using these techniques, he will explain his research into the question of how learning to perform tasks that are seemingly unrelated survival can affect the "survival of the fittest."

**Tues. 4/29 (12:00) - Effects of Learning on Coevolution
**Rebecca Wells

Department of Mathematics and Computer Science

Dickinson College

(Honor's Thesis Defense)

How can bees and flowers improve our ability to solve difficult problems using computers? For a glimpse, come to this talk! Ms. Wells will introduce techniques for using computers to simulate evolution, coevolution and learning. She will explain how she used these techniques to explore the relative problem solving power of learning combined with evolution and coevolution.

**Thurs. 4/24 (12:00) - A Link Between Mathematics and Physics: A Study of Knot Theory and Magnetic Field Line Reconnection
**Liz Bouzarth

Department of Mathematics and Computer Science

Dickinson College

(Honor's Thesis Defense)

The study of magnetic field lines is a vital area of many plasma physicists' research. Utilizing magnetic field line reconnection as inspiration for this study of knot theory led to the discovery of reconnection equivalence classes, characterized by topological helicity, an invariant for framed knots.

**Tues. 3/25 (12:30) - On Problem Solving in Ancient Arabic Mathematics
**Dr. Bernd Zimmerman

University of Jena

At least two examples from ancient Arabic mathematics (one from elementary number theory, one from elementary geometry) will be presented and discussed. You might get a feeling of problem solving methods of that epoche which might point to results several hundreds of years ahead of that time (related to Euler and Fermat).

**1/27 - AVL Trees (Adelson-Velskii and Landis)
**Dr. Hugh McGuire

Department of Computer Science

University of California Santa Barbra a

An AVL (Adelson-Velskii and Landis) tree is a binary search tree with a balance condition. The balance condition is clear and relatively easy to maintain, and it ensures that the depth of the tree is O (log N), where N is the number of items stored in the tree. Issues of the implementation in C++ will also be noted from a course on intermediate data structures and algorithms, using the textbook "Data Structures & Algorithm Analysis in C++" by Mark Allen Weis.

**Thurs. 1/23 (12:00) - Constraint Satisfaction: Computing at the Speed of Stupidity
**Dr. Tim Wahls

Department of Computer Science

Hood College

Computers are extremely fast, but programming computers to do something useful is extremely slow. If programming is considered as teaching a computer to solve a problem, then the computer is a very poor student. The programmer/teacher must break the process of finding the solution down into tiny steps, and then tell the computer exactly what order to do the steps in. In this talk, we examine constraint satisfaction as a technique for increasing the "intelligence" of the computer. Using constraint satisfaction allows the programmer to just describe the solution to the computer, rather than giving detailed step-by-step instructions for finding the solution. This approach greatly reduces the time needed to write programs, and the time needed to run such programs is often still acceptable.

**Tues. 1/21 (12:00) - Searching in Uncertain Places; An Overview of Evolutionary Computing
**Dr. Andy Novobilski

Department of Computer Science

University of Tennessee at Chattanooga

Dr. Novobilski will discuss data mining from the perspective of identifying Bayesian Network Models for forecasting future outcomes from existing data. A brief introduction to evolutionary computing (genetic algorithms and genetic programming) is given with examples of their use in forecasting time series data. Also discussed are future directions and opportunities for collaborative research in the area of Biomedical Informatics.