Vladimir Temlyakov



Education
Habilitation  Mathematics  Steklov Institute of Mathematics, Russia  1981 
Ph.D.  Mathematics  Steklov Institute of Mathematics, Russia  1978 
M.S.  Mathematics  Moscow Institute of Physics and Technology, Russia  1976 
Experience
2007 – Present  Carolina Distinguished Professor  Department of Mathematics, University of South Carolina 
1992 – Present  Associate Member  Steklov Institute of Mathematics, Russia 
1995 – 2007  Professor  Department of Mathematics, University of South Carolina 
1992 – 1995  Associate Professor  Department of Mathematics, University of South Carolina 
1986 – 1992  Leading Scientist  Steklov Institute of Mathematics, Russia 
1986 – 1992  Professor (parttime)  Moscow Institute of Physics and Technology, Russia 
1979 – 1986  Assistant/Associate Professor (parttime)  Moscow Institute of Physics and Technology, Russia 
1978 – 1986  Scientist  Steklov Institute of Mathematics, Russia 
Research
Research Interests
 Greedy approximation. I continue to actively work in both greedy approximation with respect to bases and greedy approximation with respect to redundant systems (dictionaries). Lebesguetype inequalities for greedy approximation are of special interest.
 Compressed sensing. Exact and approximate recovery by l_{1} minimization and greedy methods. Incoherent systems. Compressed sensing in Banach spaces.
 Learning theory. Mathematical aspects of supervised learning. Application of approximation theory and greedy approximation in regression and classification problems.
 Numerical integration. Multivariate numerical integration. Discrepancy theory. Fibonacci cubature formulas. Sparse grids.
Current Projects

Greedy expansions in Hilbert Spaces. Jointly with J. Nelson we study rate of convergence of expansions of elements in a Hilbert space $H$ into series with regard to a given dictionary $D$. Our primary goal is to study representations of an element $f\in H$ by a series
$f\sim \displaystyle\sum\limits _ {j=1}^\infty c _ j(f)g _ j(f), \quad g _ j(f) \in D.$
In building such a representation we should construct two sequences: $\{g _ j(f)\} _ {j=1}^\infty$ and $\{c _ j(f)\} _ {j=1}^\infty$. We consider construction of $\{g _ j(f)\} _ {j=1}^\infty$ that is based on ideas used in greedytype nonlinear approximation. This explains the use of the term greedy expansion.
Known algorithms provide rate of convergence of greedy expansions of $f\in A _ 1(D)$ slower than $m^{1/4}$. It is an interesting open problem: What is the best possible rate of convergence of greedy expansions for $f\in A _ 1(D)$? Our recent qualitative result is that the best possible rate of convergence in the above problem is faster than $m^{1/4}$. We proved that it is faster than $m^{2/7}$.

Lebesguetype inequalities for greedy approximation in Banach spaces. Jointly with D. Savu we study sparse representations and sparse approximations with respect to incoherent dictionaries. We address the problem of designing and analyzing greedy methods of approximation. A key question in this regard is: How to measure efficiency of a specific algorithm? Answering this question we prove the Lebesguetype inequalities for algorithms under consideration. A very important new ingredient is that we perform our analysis in a Banach space instead of a Hilbert space. It is known that in many numerical problems users are satisfied with a Hilbert space setting and do not consider a more general setting in a Banach space. There are known arguments that justify interest in Banach spaces. We give one more argument in favor of consideration of greedy approximation in Banach spaces. We introduce a concept of $M$coherent dictionary in a Banach space which is a generalization of the corresponding concept in a Hilbert space. We analyze the QuasiOrthogonal Greedy Algorithm (QOGA), which is a generalization of the Orthogonal Greedy Algorithm (Orthogonal Matching Pursuit) for Banach spaces. It is known that the QOGA recovers exactly $S$sparse signals after $S$ iterations provided $S<(1+1/M)/2$. This result is well known for the Orthogonal Greedy Algorithm in Hilbert spaces. The following question is of great importance: Are there dictionaries in $\mathbb{R}^n$ such that their coherence in $\ell _ p^n$ is less than their coherence in $\ell _ 2^n$ for some $p\in (1,\infty)$? We show that the answer to the above question is "yes". Thus, for such dictionaries, replacing the Hilbert space $\ell _ 2^n$ by a Banach space $\ell _ p^n$ we improve an upper bound for sparsity that guarantees an exact recovery of a signal.
BACK TO TOP
Teaching Activities
Previous Courses
 MATH 241  Vector Calculus
 MATH 550  Vector Analysis
 MATH 703  Analysis I
 MATH 704  Analysis II
 MATH 725  Approximation Theory
 MATH 728C  Selected Topics: Compressed Sensing
 MATH 728G  Selected Topics: Greedy Approximation
 MATH 728L  Selected Topics: Learning Theory
 MATH 728N  Selected Topics: Numerical Integration
 MATH 729  Nonlinear Approximation Theory
BACK TO TOP
Honors and Other Special Scientific Recognition
 USC Educational Foundation Award for Research in Science, Mathematics and Engineering, 2003.
 Invited Hour Speaker, Foundations of Computational Mathematics, Santander, Spain, July 8, 2005.
 Invited Hour Speaker, Curves and Surfaces, Avignon, France, June 29, 2006.
 Invited fortyfive minutes lecture in the section Analysis of the International Congress of Mathematicians, Madrid, Spain, August 2006.
 Carolina Distinguished Professor, August 15, 2007.
BACK TO TOP
5 Selected Publications
 D. Donoho, M. Elad, and V.N. Temlyakov, On the Lebesgue type inequalities for greedy approximation, J. Approximation Theory 147 (2007), 185195.
 V.N. Temlyakov, Greedy approximation, Acta Numerica (2008), Cambridge University Press, 235409.
 V.N. Temlyakov, Mingrui Yang, and Peixin Ye, Greedy approximation with regard to nongreedy bases, Adv. Comput. Math. 34 (2011), 319337.
 B.S. Kashin and V.N. Temlyakov, A remark on compressed sensing, Matem. Zametki 13 (2007), 7186.
 V.N. Temlyakov, Greedy Approximation, Cambridge University Press, Cambridge, 2011.
BACK TO TOP
IMI Preprints and Seminars
Go to the list of 64 preprints and 13 seminars by Vladimir Temlyakov.BACK TO TOP