Lecture Previews

Masahiro Fujita

Monday at 2 p.m.

Gates Hillman Complex 6501

Masahiro Fujita, a professor in the VLSI Design and Education Center at the University of Tokyo in Japan, will deliver a seminar titled “Tsunami Simulation on FPGA and its Analysis Based on Statistical Model Checking.”

The lecture will discuss the ways in which the simulation of tsunamis can be accelerated using field-programmable gate arrays (FPGA) and graphics processing units (GPU).

Because of their different memory structures, FPGA and GPU deal with variables in different ways. The lecture will also cover error checking of the bounds of tsunami simulations.

Fujita received his Ph.D. from the University of Tokyo. He is the co-author of seven books and the author of over 100 publications.

Biomedical Engineering Seminar Series: Randy Bartels

Tuesday at 4:30 p.m.

Breed Hall (Margaret Morrison 103)

Randy Bartels, a professor of electrical engineering at Colorado State University, will give a talk titled “Advances in Speed and Resolution in Nonlinear and Fluorescent Optical Microscopies.”

The lecture will discuss the recent developments in Bartels’ lab, which has generated three-dimensional image acquisitions in approximately one millisecond of exposure.

Bartles says that the results could have an effect on activities such as embryonic heart imaging and three-dimensional imaging of neural circuitry.

Bartels will also discuss the work his lab has completed in enhancing the resolution of three-dimensional images.

Algorithmic Economics Seminar: Michael Kearns

Tuesday at noon

Newell-Simon Hall 3305

Michael Kearns, a professor of computer and information science at the University of Pennsylvania, will discuss experiments that have been performed with human subjects in microeconomic trading networks.

In the experiments, the players were divided into two groups with mutual interests in trade, and then arranged in networks that created potential asymmetries in negotiating power. Players could only trade with their immediate neighbors. This model has a detailed equilibrium theory in which variation in individual wealth is directly related to global network structure. Kearns will discuss his findings, as well as several other analyses.

Machine Learning Special Seminars: Benjamin Van Durme

Thursday at 1:30 p.m.

Newell-Simon Hall 1305

Benjamin Van Durme, a professor of computer science at Johns Hopkins University, will deliver a seminar titled “Simple Randomized Algorithms for Large-scale HLT.”

Van Durme will give an overview of two methods for dealing with large-scale language data with streaming and randomized algorithms: Bloom filters and locality-sensitive hashing. He will discuss how the methods have been used within the HLT computer language, such as for counting over large key sets in limited memory. Van Durme, a former Carnegie Mellon graduate student, specializes in natural language processing, streaming and randomized algorithms, and psycholinguistics.