Department of Physics
University of Washington/Seattle

MS Degree Program:
Large-Scale Data Acquisition and Analysis

Introduction

Contemporary scientific research frequently involves "data mining", using huge volumes of raw data which must be distilled, analyzed and displayed using complex, sophisticated and constantly advancing techniques. Students completing this MS degree program will acquire skills and experience enabling them to lead data acquisition and analysis efforts in a wide variety of basic and applied research environments. Courses cover a variety of hardware and software issues, but are less concerned with specific contemporary solutions than with fundamental techniques which will evolve in the near future. The program will use a "case study" approach, in which students have an opportunity to actively participate in forefront research projects where the acquisition and analysis of large-scale databases plays a crucial role. Course instructors have direct experience and active involvement in ongoing applications of the subject matter, in areas such as particle physics, optical, radio and particle astrophysics, acoustical oceanography, oceanographic environmental studies, geophysical and seismological research.

Prerequisites and Degree Requirements

Applicants are expected to have the following prerequisites:

  1. A BS degree with major in a physical science, an engineering discipline, mathematics or computer science.
  2. Mathematical preparation at least including courses in calculus, elementary differential equations, complex variables, matrices.
  3. Competency in at least one high-level programming language (C, C++, Fortran, Perl, or equivalent).
  4. Elementary familiarity with Unix/Linux operating systems.

Program requirements:

  1. Complete the 9 credits of core courses, plus 9 credits of electives selected from the list below.
  2. Complete a total of 9 additional credits of electives, which may include any courses numbered 400 and above in Physics, Astronomy, Oceanography or Geophysics.
  3. Complete an Independent Study project under the supervision of a faculty member in Physics, Astronomy, Oceanography or Geophysics. Present a written report, and pass an oral examination on the independent study project before a committee of at least two graduate faculty members, including the supervising faculty member.

Courses offered:

The three core courses are taken by all students; all others are electives. All courses are 3 credits and meet 3 hr weekly, except for two core courses (Data Acquisition Techniques and Elementary Large-Scale Data Handling), which are lecture/lab courses and meet 6 hr/week, and the Seminar, which is 1 hour weekly.

Hands-on laboratory course in which students learn to use contemporary high-bandwidth DAQ hardware (VMEBus, FireWire, USB) to interface and control research equipment generating high-volume data.

Fundamentals of numerical and statistical analysis of data: numerical precision, pseudo-random numbers, probability distributions, elementary statistics, including applications of Binomial, Poisson and Normal distributions. Fourier methods, filtering and wavelet techniques.

Methods and algorithms for multi-level decision making and data filtration, and the storage, reduction, retrieval and archiving of large-scale databases, with emphasis on case studies, commonly used software packages, and hands-on experience working with databases from ongoing research projects in Physics, Astronomy, Oceanography and Geophysics.

Hypothesis testing and parameter estimation techniques. Linear and non-linear least squares, maximum likelihood, robust estimation, neo-Bayesian techniques. Determination of confidence levels and confidence regions, with case studies using data from ongoing research projects in Physics, Astronomy, Oceanography and Geophysics.

Modeling physical phenomena and experimental apparatus. Pseudorandom number generators, common algorithms for specific distributions, commonly used modeling software. Students will have an opportunity to analyze and contribute to the development of complex monte carlo simulations used in forefront research projects.

Contemporary image analysis techniques, with emphasis on practical applications in physical science and engineering research and development, rather than robotics or virtual reality. Filters, histograms, edge-finding, boundary-tracing, area computation, counting, false color.

Advanced applications of Fourier methods, wavelet and related methods, Kalman techniques. Acoustical oceanography data will be used as an example. Students will participate in analysis of large scale data from ongoing research.

Students will be introduced to a variety of commonly used data visualization and manipulation tools, such as IDL, PAW, IRAF, and gnuplot, and will practice using them to display data. Examples of useful and effective data displays from current research, and the code used to generate them, will be reviewed in detail.

Techniques for rapidly processing image data and using results to control and direct hardware operations. Emphasis will be on applications in basic research in physical science and engineering R&D rather than in conventional robotics.

Students will be guided through a set of increasingly challenging examples drawn from basic and applied research, learning the numerical and programming techniques required, and then learning to code the software needed to perform the appropriate analyses.

Weekly 1-hr seminar in which guest speakers present examples of applications of high rate DAQ and large-scale data analysis taken from from academic, industrial and government R&D projects. Emphasis on student interaction and participation.

Variable credit, according to student needs. By arrangement with faculty members in participating departments. Students work on a research project of their choice, under the supervision of a faculty member, often in collaboration with other graduate students and postdoctoral researchers.

 

 Last updated: 02/22/2000, RJW