CONTACT: David F. Salisbury, News Service (415) 725-1944;
Stanford statistician elected to National Academy of Sciences
Statistics Professor David L. Donoho has been elected to the National Academy of Sciences.
On Tuesday, April 28, the academy, a private organization of scientists and engineers established in 1863 by an act of Congress, named 60 Americans and 15 foreign associates as new members "in recognition of their distinguished and continuing achievements in original research." Election to the academy is considered one of the highest honors a scientist can achieve.
This brings the total number of Stanford faculty serving on the academy to 107, plus an additional four affiliated with the Hoover Institution.
Donoho's field is theoretical statistics. He studies the best way to analyze and process random data to reach conclusions or obtain an undistorted picture or signal despite noise and missing data. He uses statistical theory to gain a precise idea how much data are needed to achieve certain goals and to suggest optimal ways to reach those goals. He has done ground-breaking research in this field, which has been applied in a number of different areas ranging from medical imaging to seismology and astronomy.
One of his earliest interests was in the problem of ≥blind signal processing≤≠ recovering a signal that has been blurred in an unknown fashion. This topic is of interest today in cellular phone and mobile communications. He later worked to develop ≥robust≤ statistics ≠ methods that are insensitive to drastic errors contaminating only a small part of the data. This is particularly useful with data sets that have been automatically collected by computer. These often contain a lot of bad data and are so large that they are impossible to clean up by hand.
The same approach can be applied with a twist to solve signal recovery problems that involve extracting meaningful information from noisy data. In this case the data that lie outside the norm are the desired facts for astronomers searching for clues to the internal structure of a distant object or geologists examining seismic data for the blip that may mean the presence of oil.
Recently Donoho has been interested in how to use ideas like wavelets and other novel mathematical tools to help scientists get sharper signals and images. This has allowed him to chart new areas for statistical research.
He has done joint work with magnetic resonance spectroscopists to determine the structure of large molecules. And he has assisted scientists at Chevron in developing a method for compressing data produced by ocean seismic surveys that has become an industry standard.
Donoho also has been involved in the development of software tools that scientists can use to display complex, multi-dimensional data. These include the MacSpin program, the Interactive Statistical Package language, and WaveLab, which is available for free over the Internet.
Donoho earned his bachelor's degree from Princeton in 1978 and his doctorate from Harvard in 1984. He joined the faculty at the University of California-Berkeley in 1984 and moved across the Bay to Stanford in 1990.
In 1991 he received a MacArthur Fellowship. In 1992 he was elected to the American Academy of Arts and Sciences and in 1994 he received the Presidents' Award from the Committee of Presidents of Statistical Societies.
By David F. Salisbury