JEFFREY S. VITTERProfessor of Computer Science and Engineering
Email: jsv "at" tamu.edu
M.B.A. Duke University, 2002
NEW ITEM:
Interview video as part of the Purdue Oral History
Archives.
NEW ITEM: Interview video as part of the ACM SIGMOD Distinguished Profiles in Databases series. NEW ITEM: Algorithms and Data Structures for External Memory, by Jeffrey Scott Vitter, Series on Foundations and Trends in Theoretical Computer Science, now Publishers, Hanover, MA, 2008. NEW ITEM: Dr. Vitter and his former PhD student Dr. Min Wang are recipients of the 2009 Test of Time Award from the Association for Computing Machinery Special Interest Group on Management of Data (ACM SIGMOD) for their 1999 SIGMOD paper, "Approximate Computation of Multidimensional Aggregates of Sparse Data Using Wavelets." The award acknowledges the paper of the SIGMOD conference held 10 years earlier that has had the most impact in terms of research, products, and methodology over the past decade. BIOGRAPHYJeff Vitter is professor of computer science and engineering at Texas A&M University in College Station, Texas. From 2008 to 2009, he served as provost and executive vice president for academics at Texas A&M, where he had the responsibility of chief academic officer for a university of over 48,000 students and 2,700 faculty members, including the Mays Business School, Dwight Look College of Engineering, George Bush School of Government and Public Service, and the Colleges of Agriculture and Life Sciences, Architecture, Education and Human Development, Geosciences, Liberal Arts, Science, and Veterinary Medicine and Biomedical Sciences. In addition, he oversaw the academic mission of the Texas A&M University campus in Doha, Qatar. In collaboration with deans and faculty, Dr. Vitter successfully launched a number of important recruiting efforts and far-reaching initiatives, including those dealing with faculty start-up allocations, sustaining multidisciplinary initiatives, balanced scorecard reviews, and diversity. Most significantly, he initiated and led the campus-wide development of an Academic Master Plan — with Roadmaps in Teaching-Learning, Research, and Engagement, along with overarching enablers — that will guide Texas A&M to the destination set out 10 years ago in Vision 2020 as a top-10 comprehensive public university. From 2002 to 2008, Dr. Vitter served as the Frederick L. Hovde Dean of the College of Science and Professor of Computer Science at Purdue University in West Lafayette, Indiana. As dean, he was the chief academic officer and administrator of the College of Science. In approximate terms, the College of Science comprised 325 faculty members, 550 staff members, 1,000 graduate students, and 2,800 undergraduate majors, with a total annual budget of $130 million. The courses offered by the College accounted for about one-fourth of the University's 1 million student credit hours. Dr. Vitter was responsible for overseeing the discovery, learning, engagement, and diversity activities of the College of Science's seven academic departments: Biological Sciences, Chemistry, Computer Sciences, Earth & Atmospheric Sciences, Mathematics, Physics, and Statistics. Dr. Vitter led the collaborative development of two strategic plans for the College, which established a dual focus of excellence in the core departments as well as in multidisciplinary collaborations. The College grew by 60 faculty members during his tenure, several hired under the innovative COALESCE faculty hiring program targeting College-wide priorities. He also launched a comprehensive study of the undergraduate program, which resulted in an innovative outcomes-based College curriculum approved by the faculty and implemented in 2007. Several programs in the College are ranked among the very best nationally. A summary of strategic initiatives and accomplishments can be found in his expanded biography or in his curriculum vitæ. From 1993 to 2002, Dr. Vitter held a distinguished professorship at Duke University in Durham, North Carolina, where he was the Gilbert, Louis, and Edward Lehrman Professor of Computer Science. He served at Duke as chair of the Department of Computer Science in the College of Arts and Sciences from 1993–2001 and as co-director and a founding member of Duke's Center for Geometric and Biological Computing from 1997–2002. As chair, he led the Department to significant improvements in stature — characterized by a top-20 ranking, stellar faculty hires, dynamic strategic plans, a departmental culture of inclusiveness, comprehensive curriculum redesign, administrative reorganization, substantial increases in both the undergraduate and graduate programs, creation of a successful industry partners program, and a rise in sponsored research expenditures to 250% of initial level. Previously from 1980–1993, he progressed through the faculty ranks and served in various leadership roles at Brown University in Providence, Rhode Island. His educational degrees include a B.S. with highest honors in Mathematics in 1977 from the University of Notre Dame in Notre Dame, Indiana; a Ph.D.in Computer Science under Don Knuth in 1980 from Stanford University in Stanford, California; and an M.B.A. in 2002 from the Fuqua School of Business at Duke University. His home town is New Orleans, Louisiana (as everyone who knows him knows!). Dr. Vitter serves on the Board of Advisors for the School of Science and Engineering at Tulane University in New Orleans and the Visiting Committee of the Institut National de Recherche en Informatique et en Automatique (INRIA) in Rocquencourt, France. From 2000–2009, Dr. Vitter served on the Board of Directors of the Computing Research Association (CRA), where he continues to co-chair the Government Affairs Committee. He has served as Chair of ACM SIGACT, the Special Interest Group on Algorithms and Computation Theory of the world's largest computer professional organization, the Association for Computing Machinery. He has served on the Executive Council of the EATCS (European Association for Theoretical Computer Science), as well as on various review committees. Sabbatical sites have included Mathematical Sciences Research Institute in Berkeley; INRIA in Rocquencourt, France; Ecole Normale Supérieure in Paris; Bell Laboratories in Murray Hill, New Jersey; and INRIA in Sophia Antipolis, France. Dr. Vitter has been named a Guggenheim Foundation fellow, a fellow of the Association for Computing Machinery (ACM), a fellow of the Institute of Electrical and Electronics Engineers (IEEE), a National Science Foundation Presidential Young Investigator, a Fulbright Scholar, and an IBM Faculty Development Awardee. He has over 280 book, journal, conference, and patent publications reflecting his research interests described below. His Google Scholar h-index is 54. His book Algorithms and Data Structures for External Memory (now Publishers, 2008) covers the I/O field he helped found. He coauthored the books Efficient Algorithms for MPEG Video Compression (Wiley & Sons, 2002) and Design and Analysis of Coalesced Hashing (Oxford University Press, 1987). He is coeditor of the collections External Memory Algorithms and Algorithm Engineering. His editorial board memberships have included Algorithmica, Communications of the ACM, IEEE Transactions on Computers, Theory of Computing Systems (formerly Mathematical Systems Theory: An International Journal on Mathematical Computing Theory), and SIAM Journal on Computing; in addition, he has edited several special issues. He has consulted widely and is co-holder of patents in the areas of external sorting, parallel I/O, prediction, and approximate data structures. He proposed the concept and participated in the design of what has become the Purdue University Research Expertise database (PURE) and the Indiana Database for University Research Expertise (INDURE), www.indure.org. RESEARCH INTERESTSIn his research, Jeff Vitter seeks to exploit the rich interdependence between computing theory and practice. Dr. Vitter has pioneered the development of several important subfields dealing with massive data. He is perhaps best known as a founder of the field of external memory algorithms, which focuses on alleviating the I/O bottleneck between fast internal memory and slow external storage (such as disk). The goal is to design algorithms that exploit locality and parallelism in order to reduce I/O costs, which is important in a variety of data-intensive applications. His recent book, Algorithms and Data Structures for External Memory, serves as a reference for the field. He has developed efficient I/O algorithms for external memory and hierarchical memory in several domains, including geographic information systems (GIS), databases, computational science and engineering, sorting, text and string indexing, matrix computations, graph traversal, range search, data mining, and a variety of computational geometry and combinatorial problems. His approach based upon duality for utilizing parallel disks, in which communication with each disk can occur simultaneously, has led to state-of-the-art methods for sorting. He has contributed to algorithm engineering via the TPIE system (Transparent Parallel I/O programming Environment). A second key aspect of massive data where Dr. Vitter plays a leadership role is compressed data structures, where the goal is to operate directly upon compressed representations of data, yet still achieve fast search time. The wavelet tree data structure he co-developed (not to confuse with wavelets discussed two paragraphs below) is a practical and elegant structure for coding sequences of characters from a multicharacter alphabet. It has become a key technique in modern text indexing and compression, with numerous applications in science, engineering, and the liberal arts. Until about 12 years ago, fast data structures for text indexing (such as suffix trees and suffix arrays) required several times more space than the data being indexed! Based upon a recursive decomposition of the suffix array, Dr. Vitter and colleagues invented the compressed suffix array, which is substantially smaller — the first index ever provably shown to use only linear space, and then later the first ever whose size per character was provably shown to be asymptotic (i.e., with constant of proportionality 1) to the higher-order entropy of the text. The index can even reconstruct the original text in a random access manner, and thus the original text can be discarded. The net effect is that the text can be completely replaced by an index structure that has the size of compressed text but can be queried quickly. In a third aspect of massive data, Dr. Vitter is well known for his fundamental work on the design and analysis of data compression and arithmetic coding methods for text, images, and video. A provably efficient algorithm for adaptive Huffman coding bears his name. With a former student, Dr. Vitter invented the FELICS algorithm for lossless image compression; it was subsequently implemented in hardware as part of NASA's Mars Reconnaissance Orbiter. It introduced a low-cost prediction framework that led to algorithms ultimately adopted into the Lossless JPEG standard. In video compression, Dr. Vitter and his group proposed the paradigm of minimizing the combined measure of rate plus distortion for significantly improved motion estimation coding; this rate-distortion optimization has since been incorporated into the H.264/MPEG-4 AVC standard's reference encoder, used widely in the computing and communications industry. Fourth, Dr. Vitter and collaborators were the first in the database and systems communities to apply wavelets and compression techniques as key tools for summarizing, approximating, and predicting data. Wavelets have since become heavily used in database optimization, data warehousing, data streams, image processing, and data mining. For his work on wavelets for approximating high-dimensional aggregates, he and his coauthor were the recipients of the 2009 ACM SIGMOD Test of Time Award, which recognizes its paper from 10 years earlier that has had the most impact in the following decade in terms of research, products, and methodology. Dr. Vitter has co-developed novel machine learning and prediction mechanisms based upon data compression, using the principle that the more compressible a sequence is, the more predictable it is. His universal prediction algorithms for online prefetching are provably asymptotically optimal (i.e., with constant of proportionality 1). The methods predict as well as special-purpose methods tuned to the characteristics of the sequence of page requests. His learning work includes algorithms for prefetching, caching, data streams, database query optimization, data mining, and power management in mobile computers. Beginning with his thesis on coalesced hashing, a search method used widely in practice, Dr. Vitter has made many contributions to the analysis of algorithms, using mathematical analysis and asymptotics to derive precise estimates for resource requirements. He has also done much work involving randomized, parallel, and incremental algorithms for a variety of problems in computational geometry, combinatorial optimization, graphics, random sampling, and random variate generation. SELECTED PUBLICATIONS
ONLINE PUBLICATION LIBRARYDr. Vitter has over 280 book, journal, conference, and patent publications reflecting the areas of interest described above. His Google Scholar h-index is 54. Several of his recent publications are available electronically via his online publication library. CURRICULUM VITÆDr. Vitter's curriculum vitæ contains more complete information, including a full list of publications and funding. AWARDS
EXPERIENCE
PROFESSIONAL ACTIVITIES AND SERVICE
PERSONAL LINKS
|
| ||||
| ||||