KEIVAN MONFARED
Mathematics and Data Science


Publications

Under Review

  1. research/Community_structure_detection_and_evaluation_during_preictal_and_postictal_hippocampal_depth_recordings (2018+) Community structure detection and evaluation during preictal and postictal hippocampal depth recordings
    Keivan Hassani Monfared, Kris Vasudevan, Jordan. S. Farrell, and G. Cam Teskey: PDF arXiv
    We show how anti-correlation between neurons firing rates forms communities of neurons that influence the behaviour during an eppleptic seizure.
  2. research/Inverse_spectral_problems_for_linked_vibrating_systems_and_structured_matrix_polynomials (2018+) Inverse spectral problems for linked vibrating systems and structured matrix polynomials
    Keivan Hassani Monfared and Peter Lancaster: PDF arXiv
    We show that for a given set of \( nk \) distinct real numbers \( \Lambda \), and \( k \) graphs on \( n \) nodes, \( G_0, G_1,\cdots,G_{k-1} \), there are real symmetric \( n\times n \) matrices \( A_s \), \( s=0,1,\ldots, k \) such that the matrix polynomial \( A(z) := A_k z^k + \cdots + A_1 z + A_0 \) has proper values \( \Lambda \), the graph of \( A_s \) is \( G_s \) for \( s=0,1,\ldots,k-1 \), and \( A_k \) is an arbitrary nonsingular (positive definite) diagonal matrix. When \( k=2 \), this solves a physically significant inverse eigenvalue problem for linked vibrating systems.

Published

  1. research/2018_An_analog_of_Matrix_Tree_Theorem_for_signless_Laplacians (2018) An analog of Matrix Tree Theorem for signless Laplacians
    Linear Algebra and Its Applications (560) 43--55, DOI:10.1016/j.laa.2018.09.016
    Keivan Hassani Monfared and Sudipta Mallik: PDF arXiv
    A spanning tree of a graph is a connected subgraph on all vertices with the minimum number of edges. The number of spanning trees in a graph \( G \) is given by Matrix Tree Theorem in terms of principal minors of Laplacian matrix of \( G \). We show a similar combinatorial interpretation for principal minors of signless Laplacian \( Q \). We also prove that the number of odd cycles in \( G \) is less than or equal to \( \frac{\det(Q)}{4} \), where the equality holds if and only if \( G \) is a bipartite graph or an odd-unicyclic graph.
  2. research/2018_A_structured_inverse_spectrum_problem_for_infinite_graphs (2018) A structured inverse spectrum problem for infinite graphs
    Linear Algebra and Its Applications (539) 28--43, DOI:10.1016/j.laa.2017.10.025
    Keivan Hassani Monfared and Ehssan Khanmohammadi: PDF arXiv
    In this paper it is shown that for a given infinite graph \( G \) on countably many vertices, and a compact, infinite set of real numbers \( \Lambda \) there is a real symmetric matrix \( A \) whose graph is \( G \) and its spectrum is \( \Lambda \). Moreover, the set of limit points of \( \Lambda \) equals the essential spectrum of \( A \), and the isolated points of \( \Lambda \) are eigenvalues of \( A \) with multiplicity one. It is also shown that any two such matrices constructed by our method are approximately unitarily equivalent.
  3. research/2017_Existence_of_a_not_necessarily_symmetric_matrix_with_given_distinct_eigenvalues_and_graph (2017) Existence of a not necessarily symmetric matrix with given distinct eigenvalues and graph
    Linear Algebra and Its Applications (527) 1--11, DOI:10.1016/j.laa.2017.04.006
    Keivan Hassani Monfared: PDF arXiv
    For given distinct numbers \( \lambda_1 \pm \mu_1 \rm{i}, \lambda_2 \pm \mu_2 \rm{i}, \ldots, \lambda_k \pm \mu_k \rm{i} \in \mathbb{C} \setminus \mathbb{R} \) and \( \gamma_1, \gamma_2, \ldots, \gamma_l \in \mathbb{R} \), and a given graph \( G \) with a matching of size at least \( k \), we will show that there is a real matrix whose eigenvalues are the given numbers and its graph is \( G \). In particular, this implies that any real matrix with distinct eigenvalues is similar to a real, irreducible, tridiagonal matrix.
  4. research/2016_The_nowhere-zero_eigenbasis_problem_for_a_graph (2016) The nowhere-zero eigenbasis problem for a graph
    Linear Algebra and Its Applications (505) 296--312 , DOI:10.1016/j.laa.2016.04.028
    Keivan Hassani Monfared and Bryan L. Shader: PDF
    Using previous results and methods, it is shown that for any connected graph \( G \) on \( n \) vertices and a set of \( n \) distinct real numbers \( \Lambda \), there is an \( n\times n \) real symmetric matrix \( A \) whose graph is \( G \), its spectrum is \( \Lambda \), and none of the eigenvectors of \( A \) have a zero entry.
  5. research/2016_On_the_principal_permanent_rank_characteristic_sequences_of_graphs_and_digraphs (2016) On the principal permanent rank characteristic sequences of graphs and digraphs
    Electronic Journal of Linear Algebra (31) 187--199, DOI:10.13001/1081-3810.3117
    Keivan Hassani Monfared, Paul Horn, Franklin H. J. Kenter, Kathleen Nowak, John Sinkovic, and Josh Tobin: PDF arXiv
    Given an \( n\times n \) real matrix \( A \), the principal characteristic perrank sequence of \( A \) is defined as \( r_0, r_1, r_2, \ldots, r_n \), where for \( k \geq 1 \), \( r_k=1 \) iff \( A \) has a principal submatrix of size \( k \) with nonzero permanent, and \( r_0 = 1 \) iff \( A \) has a zero on its main diagonal. We study the following inverse problem: given a sequence \( r_0 r_1 \cdots r_n \) of zeros and ones, does there exist an \( n \times n \) real matrix which achieves this sequence? As a result, we characterize all the sequences corresponding to (symmetric) entry-wise nonnegative matrices, and provide some results for the skew-symmetric case.
  6. research/2016_Spectral_characterization_of_matchings_in_graphs (2016) Spectral characterization of matchings in graphs
    Linear Algebra and Its Applications (496) 407--419, DOI:10.1016/j.laa.2016.02.004
    Keivan Hassani Monfared and Sudipta Mallik: PDF arXiv
    A spectral characterization of the matching number (the size of a maximum matching) of a graph is given. More precisely, it is shown that the graphs \( G \) of order \( n \) whose matching number is \( k \) are precisely those graphs with the maximum skew rank \( 2k \) such that for any given set of \( k \) distinct nonzero purely imaginary numbers there is a real skew-symmetric matrix \( A \) with graph \( G \) whose spectrum consists of the given \( k \) numbers, their conjugate pairs and \( n-2k \) zeros.
  7. research/2015_The_lambda-tau_structured_inverse_eigenvalue_problem (2015) The \( \lambda \)-\( \tau \) structured inverse eigenvalue problem
    Linear and Multilinear Algebra (63) 2275--2300, DOI:10.1080/03081087.2014.1003527
    Keivan Hassani Monfared and Bryan L. Shader: PDF
    For a matrix \( A \) let \( A(\{r,s\}) \) denote the principal submatrix obtained form \( A \) by removing rows and columns \( r \) and \( s \). Let \( T \) be a tree on \( n \) vertices which satisfies some necessary combinatorial restriction, \( \Lambda \) a set of \( n \) distinct real numbers, \( M \) a set of \( n-2 \) distinct real numbers, and \( r \) and \( s \) two distinct fixed vertices of \( T \). Also, assume that \( M \) and \( L \) satisfy second degree strict Cauchy interlacing inequalities and some genericity conditions. Using various combinatorial techniques it is shown that there is a real symmetric matrix \( A \) whose graph is \( T \), spectrum of \( A \) is \( \Lambda \) and spectrum of \( A(\{r,s\}) \) is \( M \). Then, using Jacobian method this result was extended to any supergraph of this family of trees. Also, some consequences of these results related to perturbations of diagonal entries of a matrix are presented.
  8. research/2015_Construction_of_real_skew-symmetric_matrices_from_interlaced_spectral_data,_and_graph (2015) Construction of real skew-symmetric matrices from interlaced spectral data, and graph
    Linear Algebra and Its Applications (471) 241--263, DOI:10.1016/j.laa.2015.01.001
    Keivan Hassani Monfared and Sudipta Mallik: PDF arXiv
    We provide analogues of the results presented in the \( 2013 \) paper for skew-symmetric matrices. This is first shown for a family of trees with nearly even branching at a vertex \( v \) (NEB trees) using various combinatorial techniques. Then the results are extended to any supergraph of such trees using the Jacobian method. Let \( G \) be a connected graph on \( n \) vertices, \( \Lambda \) a set of \( n \) distinct purely imaginary numbers which is closed under negation, and \( M \) a set of \( n-1 \) distinct purely imaginary numbers closed under negation which strictly interlaces \( \Lambda \). Also, let \( v \) be a vertex of \( G \). It is shown that if \( G \) has a spanning tree which is NEB at \( v \), then there is a real skew-symmetric matrix \( A \) whose graph is \( G \), its spectrum is \( \Lambda \), and the spectrum of \( A(v) \) is \( M \). Some properties of NEB trees are also studied.
  9. research/2013_Construction_of_matrices_with_a_given_graph_and_prescribed_interlaced_spectral_data (2013) Construction of matrices with a given graph and prescribed interlaced spectral data
    Linear Algebra and Its Applications (438) 4348--4358, DOI:10.1016/j.laa.2013.01.036
    Keivan Hassani Monfared and Bryan L. Shader: PDF
    For a matrix \( A \) let \( A(v) \) denote the principal submatrix obtained form \( A \) by removing its \( v \)-th row and column. In a 1989 paper, Duarte showed that for any given tree \( T \) on \( n \) vertices, \( \Lambda \) a set of \( n \) distinct real numbers, \( M \) a set of \( n-1 \) distinct real numbers that strictly interlaces \( \Lambda \), and \( i \) a fixed vertex of \( T \), there is a real symmetric matrix \( A \) whose graph is \( T \), spectrum of \( A \) is \( \Lambda \) and spectrum of \( A(v) \) is \( M \). In this paper we develop a method, called the Jacobian method, and show that Duarte's result holds for any connected graph.
  10. research/2010_On_the_existence_of_nowhere-zero_vectors_for_linear_transformations (2010) On the existence of nowhere-zero vectors for linear transformations
    Bulletin of the Australian Mathematical Society (82) 480--487, DOI:10.1017/S0004972710001619
    Saeed Akbari, Keivan Hassani Monfared, Mohammad Jamaali, Ehssan Khanmohammadi, and Dariush Kiani: PDF
    If for a matrix \( A \) there is a vector \( x \) such that the vectors \( x \) and \( Ax \) have no zero entries, then \( A \) is called to have the AJT property. In this paper we make further progress on the Alon-Jaeger-Tarsi (AJT) Conjecture, using combinatorial, probabilistic, and linear algebraic methods. The AJT conjecture asserts any nonsingular matrix over a field with at least 4 elements has the AJT property. It is shown that any nonzero matrix is similar to an AJT matrix, and some necessary and some sufficient conditions for a matrix to have the AJT property are given. Also, we provide a sharp bound for the number of the elements of the field in terms of the size of the matrix.

Thesis

  1. research/2014_The_Jacobian_Method_The_Art_Of_Finding_More_Needles_in_Nearby_Haystacks_(PhD_dissertation) (2014) The Jacobian Method: The Art Of Finding More Needles in Nearby Haystacks (PhD dissertation)
    Keivan Hassani Monfared, Advisor: Bryan L. Shader: PDF
    In this dissertation we develop the Jacobian method, and we give a detailed description of it, along with a geometric interpretation. We provide several examples and using the Jacobian method we solve three fundamental structured inverse eigenvalue problems including the \( \lambda \)-\( \mu \)-SIEP, the \( \lambda \)-\( \tau \)-SIEP, and the nowhere-zero eigenbasis problem for the \( \lambda \)-SIEP. The common strategy is to prove the existence of a solution for trees, then to show that these solutions are generic, and then use the Implicit Function Theorem to extend it to connected graphs. Several other important problems that can be solved using this method are mentioned, as well as two main directions for further development of the Jacobian method.
  2. research/2012_On_the_Permanent_Conjecture_(Masters_thesis) (2012) On the Permanent Conjecture (Masters thesis)
    Keivan Hassani Monfared, Advisor: Bryan L. Shader: PDF
    The permanent rank of a matrix \( A \), \( \rm{perrank}(A) \), is defined to be the size of a largest square submatrix of \( A \) with nonzero permanent. In this thesis we study some properties of the permanents and also permanent ranks of the matrices, and related problems. In particular we characterize matrices with small permanent ranks in order to study the following conjecture relating the \( \rm{perrank} \) and the \( \rm{termrank} \).
    Conjecture: For any matrix \( A \), \[ \rm{perrank}(A) \geq \Bigg\lceil \frac{\mbox{termrank}(A)}{2} \Bigg\rceil,\] and for even \( \rm{termrank} \) the equality holds if and only if up to equivalence and reduction \( A=\bigoplus \left[\begin{array}{cr}1&1\\1&-1\end{array}\right] \).

In Progress

  1. (2018+) Properties of a nonnegative reciprocal matrix and their application to AHP
    Mojtabe Eslami and Keivan Hassani Monfared
    We study the properties of a nonnegative reciprocal matrices associated with graphs and their application to {AHP}
  2. (2018+) Dissonance networks and social influence
    Omid Atabati, Mojtaba Eslami, and Keivan Hassani Monfared
    We model how a society manages conflicts between beliefs and social standings by forming a dissonance network, and then providing measures of centrality and modularity for such signed networks.

Unpublished

  1. research/2016_The_maximum_multiplicity_of_an_eigenvalue_of_symmetric_matrices_with_a_given_graph (2016) The maximum multiplicity of an eigenvalue of symmetric matrices with a given graph
    Keivan Hassani Monfared and Sudipta Mallik: PDF arXiv
    For a graph \( G \), \( M(G) \) denotes the maximum multiplicity occurring of an eigenvalue of a symmetric matrix whose zero-nonzero pattern is given by edges of \( G \). We introduce two combinatorial graph parameters \( T^-(G) \) and \( T^+(G) \) that give a lower and an upper bound for \( M(G) \) respectively, and we show that these bounds are sharp.
    It was pointed out by the referees that the bounds are not better than those given by zero forcing number of the graph.


Talks & Seminars

Invited

  1. (2019-Jul) Inverse Polynomial Spectral Problems for Graphs: Abstract
    22nd Conference of the International Linear Algebra Society (ILAS19), Rio de Janeiro, Brazil
  2. (2016-Jun) An Analog of Matrix Tree Theorem for Signless Laplacians: Abstract Slides
    Canadian Mathematical Society Summer 2019 Meeting (CMS19), Regina, SK, Canada
  3. (2017-Jul) Counting from one
    High School Math Camp at University of Calgary, Calgary, AB, Canada
  4. (2017-Jul) Counting to infinity
    High School Math Camp at University of Calgary, Calgary, AB, Canada
  5. (2016-Nov) Real life applications of calculus and linear algebra
    Student Colloquium, University of Calgary, Calgary, AB, Canada
  6. (2016-Jul) Using jacobian method to solve inverse eigenvalue problems for graphs: Abstract
    International Linear Algebra Society (ILAS 16), Leuven, Belgium
  7. (2016-Jun) Touching infinity
    Junior Math Contestants, University of Calgary, Calgary, AB, Canada
  8. (2016-Mar) Permanent ranks of matrices and generalized cycles of graphs: Abstract
    47th Southeastern International Conference on Combinatorics, Graph Theory \& Computing, Boca Raton, FL, USA
  9. (2016-Jan) On the principal permanent rank characteristic sequences of graphs: Abstract
    Joint Mathematics Meeting 2016 (JMM 16), Seattle, WA, USA
  10. (2015-Jun) Several examples on the jacobian method: Abstract
    Canadian Discrete and Algorithmic Mathematics Conference (CanaDAM 15), Saskatoon, SK, Canada
  11. (2014-Aug) The inverse principal perrank characteristic sequence problems: Abstract
    Rocky Mountain-Great Plains Graduate Research Workshop in Combinatorics (GRWC 15), Denver, CO, USA
  12. (2013-Nov) Using the jacobian method in structured inverse eigenvalue problems: Abstract
    University of Colorado, Denver, CO, USA
  13. (2008-Jun) On mathematics education
    Parviz Shahriari Scientific and Cultural Foundation, Tehran, Iran

Contributed

  1. (2019-Jun) Graph Clustering Problems Arising in Neuroscience - Results: Abstract Slides
    The 7th Canadian Discrete and Algorithmic Mathematics Conference (CanaDAM 2019), Vancouver, BC, Canada
  2. (2018-Jul) Graph partitioning problems arising in neuroscience - Questions: Abstract
    Prairie Discrete Mathematics Workshop (PDMW18), Brandon, MB, Canada
  3. (2016-May) Some inverse eigenvalue problems for graphs: Abstract
    Western Canada Linear Algebra Meeting(WCLAM), University of Manitoba, Winnipeg, MB, Canada
  4. (2016-Jan) Using the jacobian method to solve structured inverse eigenvalue problems
    Joint Mathematics Meetings, Seattle, WA, USA
  5. (2015-Oct) What do generalized cycles of a graph tell about each other?
    Research Seminars, University of Calgary, Calgary, AB, Canada
  6. (2015-Sep) How to find more solutions when you have one in hand
    Research Seminars, University of Calgary, Calgary, AB, Canada
  7. (2015-Jan) Nowhere-zero eigenbasis for a matrix with prescribed graph and spectrum: Abstract
    Joint Mathematics Meetings, San Antonio, TX, USA
  8. (2014-Oct) Building vibrating systems using linear algebra and calculus: Abstract
    Student Colloquium, Western Illinois University, Macomb, IL, USA
  9. (2014-Oct) Structured inverse eigenvalue problems: Abstract
    Colloquium, Western Illinois University, Macomb, IL, USA
  10. (2014-Apr) Skew-symmetric SIEP and the role of the jacobian method
    Algebra Combinatorics and Number Theory seminars, University of Wyoming, Laramie, WY, USA
  11. (2014-Jan) The jacobian method and structured inverse eigenvalue problems: Abstract
    Joint Mathematics Meeting, Baltimore, MD, USA
  12. (2013-Nov) On the importance of the jacobian method
    Graduate Students Seminars, University of Wyoming, Laramie, WY, USA
  13. (2013-Oct) Zonotopal algebra, an expository talk
    Algebra Combinatorics and Number Theory seminars, University of Wyoming, Laramie, WY, USA
  14. (2013-Aug) A structured inverse eigenvalue problem: Abstract
    MathFest, Hartford, CT, USA
  15. (2013-Jul) The \( \lambda \)-\( \mu \) structured inverse eigenvalue problem
    Rocky Mountain Discrete Math Days, University of Wyoming, Laramie, WY, USA
  16. (2013-Jul) The \( \lambda \)-\( \mu \) structured inverse eigenvalue problem: Abstract
    Rocky Mountain Mathematics Consortium, University of Wyoming, Laramie, WY, USA
  17. (2012-Oct) The \( \lambda \)-\( \mu \) structured inverse eigenvalue problem
    Rocky Mountain Discrete Math Days, Denver University, Denver, CO, USA
  18. (2012-Sep) A jacobian approach to some structured inverse eigenvalue problems
    Algebra Combinatorics and Number Theory seminars, University of Wyoming, Laramie, WY, USA
  19. (2012-Sep) Constructing matrices with interlacing spectral data and graph
    Midwestern Graph Theory (MIGHTY) LIII Conference, Iowa State University, Ames, IA, USA
  20. (2012-Apr) Why is the permanent rank important?
    Graduate Student Combinatorics Conference, University of Illinois, Urbana-Champaign, IL, USA
  21. (2012-Jan) On the permanent rank of matrices: Abstract
    Joint Mathematics Meeting, Boston, MA, USA
  22. (2011-Oct) Perrank v.s rank
    Rocky Mountain Discrete Math Days, University of Wyoming, Laramie, WY, USA
  23. (2011-Mar) What is the permanent?
    Graduate Students Seminars, University of Wyoming, Laramie, WY, USA
  24. (2010-May) A different approach to the hall's marriage theorem
    Graduate Students Seminars, University of Wyoming, Laramie, WY, USA
  25. (2009-Nov) A survey on the alon-jaeger-tarsi conjecture
    Algebra Combinatorics and Number Theory seminars, University of Wyoming, Laramie, WY, USA
  26. (2006--07) Biweekly math problem solving seminars
    Undergraduate Students Seminars, Amirkabir Univeersity of Technology, Tehran, Iran


Selected Presentation Slides & Videos

PhD Dissertation Defense: PDF

PhD Defence

The Jacobian method in 1 and a half napkins and a cup:

The Jacobian method in one and a half napkins and a cup

The Jacobian Method and Structured Inverse Eigenvalue Problems at JMM 14, Baltimore, MD: Video

Giving a talk on the Jacobian method at the joint math meeting 2014, Baltimore, MD

The \(\lambda\)-\(\mu\) SIEP (long): PDF

Lambda-mu SIEP

The \(\lambda\)-\(\mu\) SIEP (short): PDF

Lambda-mu SIEP

The \(\lambda\)-\(\tau\) SIEP: PDF

lambda-tau SIEP

Permanent Rank: PDF

On the permanent conjecture


Areas of Interest

Linear Algebra and Matrix Analysis
Inverse Eigenvalue Problems for Graphs
Multilinear Functions, Permanent
Combinatorics and Graph Theory
Matching Problems, Cycle Analysis
Applications in Neuroscience and Economics
Clustering, Centrality, Connectivity
Multilayer Netwroks, Dynamics on Networks
Dynamical Systems, Stability, Robustness
Signal Processing, Filtering
Data Science, Machine Learning, Deep Learning
Statistical Analysis, Regression Analysis, Hypothesis Testing


Collaborators

Saeed Akbari
Jordan Farrell
Paul Horn
Mohammad Jamaali
Franklin Kenter
Ehssan Khanmohammadi
Dariush Kiani
Peter Lancaster
Sudipta Mallik
Kathleen Nowak
Bryan Shader
John Sinkovic
Cam Teskey
Josh Tobin
Kris Vasudevan


github linkedin researchgate googlescholar arxiv wordpress twitter