Bell Laboratories
700 Mountain Avenue, 2C283
Murray Hill, New Jersey 07974
(908) 582-3869
cocteau@bell-labs.com
/who/cocteau
160 Summit Avenue
Apartment # 34
Summit, New Jersey 07901
(908) 608-0202
| 1989 - 1994 | Department of Statistics, University of California at Berkeley | |
| Dissertation Title: Extended Linear Models, Multivariat e Splines and ANOVA | ||
| 1987 - 1989 | Department of Mathematics, University of California at Berkeley | |
| 1982 - 1987 | Department of Mathematics, University of California at Davis | |
| Graduated with High Honors | ||
Polynomial splines and their tensor products are natural building
blocks for constructing finite dimensional estimates of infinite
dimensional main effects and low-order interactions, and the resulting
ANOVA decompositions provide an insightful tool for data analysis. In
order to synthesize the theoretical properties of these estimates, I
have developed the notion of an extended linear model. Many
statistical problems of theoretical and practical importance can be
effectively treated within this framework: Regression, generalized
regression, polychotomous regression, hazard regression, censored
regression, density estimation and conditional density estimation are
all examples of extended linear models. These theoretical
investigations provide valuable clues to developing sound statistical
methodology. I am particularly interested in practical applications
of multivariate spline spaces in the context of function estimation.
I have authored an extensive S chapter of functions devoted entirely
to computing with bivariate splines defined over arbitrary
triangulations in the plane. Within this environment, I have been able
to experiment with a number estimation procedures that involve
adaptively constructed spline spaces. Ultimately, these ideas will be
discussed in a monograph on extended linear modeling and polynomial
splines. Recently, I have begun to explore the use of Bayesian
computational procedures for knot placement in the context of extended
linear models. Early results indicate that this approach can greatly
improve our original work with greedy searches.
Collaborators: Jianhua Huang, Charles Kooperberg, Charles
Stone and Young Truong
Given the exponential growth of the Web, it is natural to question its
``impact.'' The meaning of this term will vary depending on the
purpose for which a given site was designed. Current software
offerings aimed at analyzing Web page statistics are typically little
more than accounting packages that calculate the number of hits per
page, and possibly a temporal or geographical load distribution. Wim
Sweldens and I have designed new techniques and supporting
visualization tools to help webmasters better understand how visitors
are using their sites: What are typical paths? How much time does the
user need to find a page? Are the most popular links easy to find?
Can users be classified by their usage of a site? A byproduct of this
effort is a model for navigation, characterizing how users move both
within and between sites.
Collaborator: Wim Sweldens
By viewing statistical modeling as a means of generating
descriptions of observed data, the Principle of Minimum Description
Length (MDL) suggests discriminating between competing models based on
the complexity of each description. This approach began with
Kolmogorov's theory of algorithmic complexity, matured in the
literature on information theory, and has recently received renewed
interest within the statistics community. My interest is in exploring
both the practical as well as the theoretical aspects of MDL as a tool
for model selection, emphasizing the rich connections between
information theory and statistics. At the boundary between these two
disciplines, we find many interesting interpretations of popular
frequentist and Bayesian procedures. MDL provides an objective
umbrella under which rather disparate approaches to statistical
modeling can co-exist and be compared.
Collaborator: Bin Yu
Semiconductor manufacturing is a complex process involving hundreds of
separate steps and lasting up to two months. Several hundred circuits
are fabricated simultaneously on a wafer, and the wafers are
themselves processed in groups called lots. A large manufacturing
plant typically processes hundreds of lots a week. Unfortunately, this
volume precludes engineers from exploiting information from all the
available data for process improvement. During the initial phase of
this research, Vijay Nair, David Friedman and I developed a series of
S-based software tools for handling probe test data as well as a
statistical strategy for routinely identifying significant patterns or
clusters in wafer map data and relating them to critical process
information. Since that point, we have widened our focus
considerably, incorporating many new data sources. Aside from some
rather aesthetic mathematics, one of the most gratifying aspects of
this project has been the fact that our techniques are now an integral
part of the routine data analysis carried out by engineers at Lucent's
manufacturing facilities in Orlando, Allentown and Madrid.
Collaborators: Vijay Nair and David James
Member of Technical Staff, Bell Laboratories
Summer 1994 - Present
Projects:
The practical use of spline spaces for function estimation
Development of the theory of Extended Linear Models
Quality improvement in IC
manufacturing
Web statistics, large data bases
Research Assistant for Professor Charles J. Stone
Summer 1992 - Spring 1994
Project: Applications of non-tensor product multivariate spline spaces
Statistical Consultant, AT&T Bell Laboratories
Summer 1992
Project: Quality improvement in semiconductor manufacturing
Sponsor: Vijay N. Nair
Statistical Consultant for Professor Kenneth
Wachter
Fall 1991 - Spring 1992
Project:
Matters of Census adjustment relating to testimony by Professors
Kenneth Wachter and Leo Breiman in
New York City, et al. versus the Department of Commerce, et al.
Summer Intern, AT&T Bell Laboratories
Summer 1991
Projects:
M-estimation and fractional factorial designs
Quality improvement in semiconductor manufacturing
Mentor: Vijay N. Nair
Research Assistant for Professor David A. Freedman
Fall 1990 - Spring 1991
Projects:
Vote dilution,
ecological regression and the neighborhood model
Preliminary analysis of the proposed
adjustment procedure for the 1990 Census
Research Assistant for Professor Charles J. Stone
Summer 1990
Project: Functional inference and experimental design
Teaching Assistant: Master's Level Introduction to Probability and Statistics
Fall 1991 - Spring 1992
University of California at Berkeley
Instructor: Professor Charles J. Stone
Received an Outstanding Teaching Award from the Graduate
Division of the University of California at Berkeley in May, 1992.
This award
is based on both student and instructor evaluations.
Teaching Assistant: Graduate Level Applied Statistics
Spring 1991
University of California at Berkeley
Instructor: Professor D. A. Freedman
Teaching Assistant: Master's Level Introduction to Probability
Fall 1990
University of California at Berkeley
Instructor: Professor Charles J. Stone
Teaching Assistant: Undergraduate Introduction to Probability
Fall 1989
University of California at Berkeley
Instructor: Professor Charles J. Stone
Workshop Leader for Upper Division Courses
in Mathematics
Fall 1987 - Spring 1989
University of California at Berkeley
Sponsor: The Minority Engineering Program
U.S. Patent No. 5,240,866:
A Method for Characterizing Defects in
Integrated Circuits August 1993
Inventors: David Friedman, Mark Hansen, James Hoyer and
Vijay Nair
System and Method for Spectroscopic Product Recognition and
Identificationfiled 1997
with Don Sun and Gordon Thomas
Method and Apparatus for Object Recognitionfiled 1998
with Harald Hess, Partha Mitra, and Gordon Thomas
Method for characterizing and visualizing patterns of
usage of a web site by network usersfiled 1998
with Wim Sweldens
System and method for providing interactive dialogue and iterataive search functions to find information
filed 2000
with Kit August, Daniel Lee and Elizabeth Shriver
Editoral service
Associate Editor, Journal of the American Statistical
Association January, 1999 - Present
Associate Editor, Technometrics October, 1998 - Present
Co-Editor, Statistical Computing and Graphics Newsletter
1996 - Present
Served as a referee for the Journal of the American Statistical
Association, Journal of the Royal Statistical Society,
Journal of Computational and Graphical Statistics,
Technometrics, IEEE Transactions on Information Theory,
Scandinavian Journal of Statistics.
Elected positions
JSM Program Chair, Statistical Computing Section of the American
Statistical Association
2001 Chair, Statistical Computing Section of the American Statistical
Association
Conference Organization
Organizer, Nonlinear Estimation and Classification
Spring 2001
Two Week MSRI Workshop
Program
committee, 12th
International Conference on Scientific and Statistical DBM
July 2000
Program committee, Workshop on Complexity and
Inference Academic year 2001-2002
DIMACS Special Year on
Computational Information Theory and Coding
Organizer, Statistical Science and the InternetJuly 1998
A mini-conference designed to explore the impact that emerging
Internet-related tools will have on the way we
communicate statistical
ideas.
M. H. Hansen and K. Wachter (1991), ``Report on CAT Experiments,"
Technical Report No. 362,
Department of Statistics, University of California
at Berkeley.
M. H. Hansen (1994), ``Extended Linear Models, Multivariate Splines
and ANOVA,'' Ph.D. Dissertation, University of California at Berkeley.
M. H. Hansen, D. James and D. Pregibon (1995),
``Tools for Spatial Data Analyses in Integrated Circuit Manufacturing,"
AT&T Technical Memorandum.
M. H. Hansen, D. James (1995),
``Extending Wafer Space: Tencor Particle Scans and Bitmaps,"
Bell Laboratories Technical Memorandum.
M. H. Hansen and V. Nair (1996),
``Process and Productivity Improvement in Semiconductor Manufacturing,"
Computing Science and Statistics, Vol. 27, pp. 3-7.
D. Friedman, M. H. Hansen, V. Nair and D. James (1997),
``Model-free estimation of some yield metrics in integrated circuit
manufacturing," with
David Friedman, Vijay Nair and David James
IEEE Transactions on Semiconductor Manufacturing, Vol. 10, No. 3.
C. Stone, M. H. Hansen, C. Kooperberg, Y. Truong (1997),
``Polynomial Splines and Their Tensor Products in Extended Linear
Modeling,"
(with discussion),
The Annals of Statistics, Vol. 25, No. 4, pp. 1371-1470.
M. H. Hansen, V. Nair and D. Friedman (1997),
``Monitoring Wafer Map Data from Integrated Circuit Fabrication Processes
for Spatially Clustered Defects," Technometrics Vol. 39, No. 3.
M. H. Hansen and D. James (1997), ``A Computing Environment for Spatial Data
Analysis in the Microelectronics Industry,'' Bell Labs Technical
Journal, Vol. 2, No. 1, pp. 114-129.
M. H. Hansen, C. Kooperberg and S. Sardy (1998), ``Triogram Models ''
JASA, Vol. 93, pp. 101-119.
O. Patterson, M. H. Hansen, and T. Henry (1999), ``Spatial Analysis of Defect
and Electrical Test Wafer Maps for Yield Limiting Defect Reduction,"
Proc. SEMATECH, May
19, 1999.
M. H. Hansen and B. Yu (1999), ``Bridging AIC and BIC: An MDL Model
Selection Criterion,'' IT Workshop on Detection, Estimation,
Classification and Imaging, February, 1999.
M. H. Hansen and B. Yu (1999), ``Assessing MDL in Wavelet Denoising and
Compression,'' ITW 1999.
M. H. Hansen and B. Yu (1999), ``Model Selection and the Principle of
Minimum Description Length,'' revision submitted to JASA.
M. H. Hansen, V. N. Nair and D. Friedman (1999), ``Process Improvement
Through the Analysis of Spatially Clustered Defects on Wafer Maps,''
submitted to JASA.
M. H. Hansen and C. Kooperberg (1999) ``Spline Adaptation in Extended
Linear Models,'' submitted to Statistical Science.
M. H. Hansen and B. Yu (2000), ``Wavelet Thresholding via MDL:
Simultaneous Denoising and Compression,'' IEEE
Tranasctions on Information Theory, Vol. 45, No. 5, pp. 1778-1788.
M. H. Hansen, V. N. Nair and J. Shi (2000), ``Statistics in Advanced
Manufacturing,'' (Year 2000 Commemorative Vignette on
Engineering and the Physical Sciences), JASA, Vol. 95, No. 451, pp. 1002-1005.
M. H. Hansen and B. Rubin (2000), ``The audiences would be the artists
and their life would be the arts," IEEE MultiMedia, Vol. 7, No. 2, pp. 6-9.
Extended Linear Modeling with Splines,
monograph on the methodological and theoretical
aspects of the use of polynomial splines and their tensor products
in multivariate function estimation,
with Charles Kooperberg, Jianhua Huang,
Charles Stone and Young Truong.
(In preparation, publication expected August, 2000.)
ORSA/TIMS Meeting, San Francisco, California
November 1992
Title: Process Improvement through Statistical Modeling and
Data Analysis: An Application to Integrated
Circuit Fabrication
Quality and Productivity Research Conference, Knoxville, Tennessee
June 1993
Title: Statistics in Advanced Manufacturing:
Process Improvement in Integrated Circuit Fabrication
1993 Joint Statistical Meetings of the ASA, San Francisco, California
August 1993
Title: Quality Improvement in Integrated Circuit Fabrication
Through Spatial Analysis of Wafer Map Data
Fall Technical Conference, Rochester, New York
October 1993
Title:
Quality Improvement in Integrated Circuit Fabrication
Spring Research Conference on Statistics in Industry and Technology,
Chapel Hill, NC
June 1994
Title: Quality and Productivity Improvement in Integrated Circuit Fabrication
Quality and Productivity Research Conference, Phoenix, Arizona
May 1995
Title: Strategies for Process and Productivity Improvement in IC Manufacturing
27th Symposium on the Interface: Computing Science and Statistics,
Philadelphia, PA
June 1995
Title: Process and Productivity Improvement in Semiconductor Manufacturing
IMS WNAR, Stanford, California
June 1995
Title: Bivariate function estimation through the use of spline
spaces defined over adaptively constructed triangulations
Invited Seminar, University of Washington Department of Statistics
February 1996
Invited Seminar, University of Illinois at
Urbana-Champaign Department of StatisticsApril 1996
1996 Joint Statistical Meetings of the ASA, Chicago, Illinois
August 1996
Title: Data Analysis and the Web
Invited Seminar, Rutgers University Department of StatisticsSeptember 1996
Invited Seminar, University of California at
Berkeley Department of StatisticsNovember 1996
Invited Seminar, Weierstrass Institute for Applied Analysis and StochasticsMarch 1997
21. Jahrestagugg der Gesellschaft für Klassifikation e.V.,
Potsdam, Germany
March 1997
Title: Spline Adaptation in Extended Linear Models
Invited Seminar, The Wharton School, University of Pennsylvania April 1997
Joint Research Conference on Statistics in Industry and Technology,
New Brunswick, NJJune 1997
Organized Invited Session: Semiconductor Yield Modeling
1997 Joint Statistical Meetings of the ASA, Anaheim, California
August 1997
Organized Invited Session: Computational Science and the Internet
1997 Joint Statistical Meetings of the ASA, Anaheim, California
August 1997
Title: Adaptive Knot Splines
Fall Technical Conference, Baltimore, Maryland October 1997
Title: Monitoring Wafer Map Data from IC Fabrication Processes for Spatially Clustered Defects
(Technometrics Invited Session)
Invited Seminar, Cornell University, Department of ORIEMarch 1998
Invited Seminar, Cornell University, Department of StatisticsApril 1998
Invited Seminar, University of Michigan, Department of
StatisticsApril 1998
1998 Joint Statistical Meetings of the ASA, Dallas, Texas
August 1998
Title: Extended Linear Modeling and the Principle of Minimum
Description Length
(Invited Talk)
4th ICSA Statistical Conference, Kunming, China
August 1998
Title: Statistics in Advanced Manufacturing
Invited Seminar, New Jersey Institute of Technology October 1998
SEMATECH Symposium on Spatial Statistics and Pattern
Recognition, Austin, TexasMay 1999
Talk: Spatial Analysis of Defect and Electrical Test
Wafer Maps for Yield Limiting Defect Reduction
Invited Half-Day Workshop (joint with V. N. Nair and D. James): Process Monitoring and Process
Improvement through Spatial Analysis of Wafermap Data
1999 Joint Statistical Meetings of the ASA, Baltimore, MD
August 1999
Title: ASA and the Web
1999 Joint Statistical Meetings of the ASA, Baltimore, MD
August 1999
Organized Invited Session: Practical Smoothing
34th Annual Conference on Information Sciences and Systems,
Princeton University
March 2000
Invited Talk: Uncovering Complexity in Data through Sound