Theo McKenzie
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Research Interests
Theo McKenzie has a total of 16 co-authors. Here are the top co-authors by number of shared publications:
- Dingding Dong
Harvard University
1 shared publication Ella Hiesmayr
ENS Lyon Lyon France
1 shared publication- Ella Hiesmayr
Department of Statistics University of California Berkeley
1 shared publication - Horng-Tzer Yau
University of Pennsylvania Harvard University
1 shared publication - Jiaoyang Huang
University of Pennsylvania Harvard University
1 shared publication John Urschel
Department of Mathematics , Massachusetts Institute of Technology, Cambridge, MA, USA
1 shared publication
Latest Publications
- 19 May 2025
- Published by Wiley in Random Structures & Algorithms
- Vol. 66 (3)
The Spectral Edge of Constant Degree Erdős–Rényi Graphs
We show that for an Erdős–Rényi graph on vertices with expected degree satisfying , the largest eigenvalues can be precisely determined by small neighborhoods around vertices of close to maximal degree. Moreover, under the added condition that , the corresponding eigenvectors are localized, in that the mass of the eigenvector decays exponentially away from the high degree vertex. This dependence on local neighborhoods implies that the edge eigenvalues converge to a Poisson point process. These theorems extend a result of Alt, Ducatez, and Knowles, who showed the same behavior for satisfying and answer a question of Guionnet. To achieve high accuracy in the constant degree regime, instead of comparing the true eigenvector to that of a tree with more regularity, we examine the eigenvector equation at each vertex in a ball to deduce localization of the eigenvector and derive a continued fraction formula for the eigenvalue. This formula can be applied to any tree and could be of independent interest, especially for rooted trees with large central degree.
- 19 May 2025
- Published by Wiley in Random Structures & Algorithms
- Vol. 66 (3)
The Spectral Edge of Constant Degree Erdős–Rényi Graphs
We show that for an Erdős–Rényi graph on vertices with expected degree satisfying , the largest eigenvalues can be precisely determined by small neighborhoods around vertices of close to maximal degree. Moreover, under the added condition that , the corresponding eigenvectors are localized, in that the mass of the eigenvector decays exponentially away from the high degree vertex. This dependence on local neighborhoods implies that the edge eigenvalues converge to a Poisson point process. These theorems extend a result of Alt, Ducatez, and Knowles, who showed the same behavior for satisfying and answer a question of Guionnet. To achieve high accuracy in the constant degree regime, instead of comparing the true eigenvector to that of a tree with more regularity, we examine the eigenvector equation at each vertex in a ball to deduce localization of the eigenvector and derive a continued fraction formula for the eigenvalue. This formula can be applied to any tree and could be of independent interest, especially for rooted trees with large central degree.
- Open Access
Arbitrary Spectral Edge of Regular Graphs
We prove that for each $d\geq 3$ and $k\geq 2$, the set of limit points of the first $k$ eigenvalues of sequences of $d$-regular graphs is \[ \{(μ_1,\dots,μ_k): d=μ_1\geq \dots\geq μ_{k}\geq2\sqrt{d-1}\}. \] The result for $k=2$ was obtained by Alon and Wei, and our result confirms a conjecture of theirs. Our proof uses an infinite random graph sampled from a distribution that generalizes the random regular graph distribution. To control the spectral behavior of this infinite object, we show that Huang and Yau's proof of Friedman's theorem bounding the second eigenvalue of a random regular graph generalizes to this model. We also bound the trace of the non-backtracking operator, as was done in Bordenave's separate proof of Friedman's theorem.Open AccessArbitrary Spectral Edge of Regular Graphs
We prove that for each $d\geq 3$ and $k\geq 2$, the set of limit points of the first $k$ eigenvalues of sequences of $d$-regular graphs is \[ \{(μ_1,\dots,μ_k): d=μ_1\geq \dots\geq μ_{k}\geq2\sqrt{d-1}\}. \] The result for $k=2$ was obtained by Alon and Wei, and our result confirms a conjecture of theirs. Our proof uses an infinite random graph sampled from a distribution that generalizes the random regular graph distribution. To control the spectral behavior of this infinite object, we show that Huang and Yau's proof of Friedman's theorem bounding the second eigenvalue of a random regular graph generalizes to this model. We also bound the trace of the non-backtracking operator, as was done in Bordenave's separate proof of Friedman's theorem. - Open Access
- 10 June 2024
- Published by Association for Computing Machinery (ACM)
- p. 788-799
Explicit Two-Sided Unique-Neighbor Expanders
We study the problem of constructing explicit sparse graphs that exhibit strong vertex expansion. Our main result is the first two-sided construction of imbalanced unique-neighbor expanders, meaning bipartite graphs where small sets contained in both the left and right bipartitions exhibit unique-neighbor expansion, along with algebraic properties relevant to constructing quantum codes. Our constructions are obtained from instantiations of the tripartite line product of a large tripartite spectral expander and a sufficiently good constant-sized unique-neighbor expander, a new graph product we defined that generalizes the line product and the routed product of previous well-known works. To analyze the vertex expansion of graphs arising from the tripartite line product, we develop a sharp characterization of subgraphs that can arise in bipartite spectral expanders, generalizing previously known results, which may be of independent interest. By picking appropriate graphs to apply our product to, we give a strongly explicit construction of an infinite family of (d1,d2)-biregular graphs (Gn)n≥ 1 (for large enough d1 and d2) where all sets S with fewer than a small constant fraction of vertices have Ω(d1· |S|) unique-neighbors (assuming d1 ≤ d2). Additionally, we can also guarantee that subsets of vertices of size up to exp(Ω(√log|V(Gn)|)) expand losslessly.
- 10 June 2024
- Published by Association for Computing Machinery (ACM)
- p. 788-799
Open AccessExplicit Two-Sided Unique-Neighbor Expanders
We study the problem of constructing explicit sparse graphs that exhibit strong vertex expansion. Our main result is the first two-sided construction of imbalanced unique-neighbor expanders, meaning bipartite graphs where small sets contained in both the left and right bipartitions exhibit unique-neighbor expansion, along with algebraic properties relevant to constructing quantum codes. Our constructions are obtained from instantiations of the tripartite line product of a large tripartite spectral expander and a sufficiently good constant-sized unique-neighbor expander, a new graph product we defined that generalizes the line product and the routed product of previous well-known works. To analyze the vertex expansion of graphs arising from the tripartite line product, we develop a sharp characterization of subgraphs that can arise in bipartite spectral expanders, generalizing previously known results, which may be of independent interest. By picking appropriate graphs to apply our product to, we give a strongly explicit construction of an infinite family of (d1,d2)-biregular graphs (Gn)n≥ 1 (for large enough d1 and d2) where all sets S with fewer than a small constant fraction of vertices have Ω(d1· |S|) unique-neighbors (assuming d1 ≤ d2). Additionally, we can also guarantee that subsets of vertices of size up to exp(Ω(√log|V(Gn)|)) expand losslessly.
- Open Access
Optimal Eigenvalue Rigidity of Random Regular Graphs
Consider the normalized adjacency matrices of random $d$-regular graphs on $N$ vertices with fixed degree $d\geq 3$, and denote the eigenvalues as $λ_1=d/\sqrt{d-1}\geq λ_2\geqλ_3\cdots\geq λ_N$. We prove that the optimal (up to an extra $N^{{\rm o}_N(1)}$ factor, where ${\rm o}_N(1)$ can be arbitrarily small) eigenvalue rigidity holds. More precisely, denote $γ_i$ as the classical location of the $i$-th eigenvalue under the Kesten-Mckay law in decreasing order. Then with probability $1-N^{-1+{\rm o}_N(1)}$, \begin{align*} |λ_i-γ_i|\leq \frac{N^{{\rm o}_N(1)}}{N^{2/3} (\min\{i,N-i+1\})^{1/3}},\quad \text{ for all } i\in \{2,3,\cdots,N\}. \end{align*} In particular, the fluctuations of extreme eigenvalues are bounded by $N^{-2/3+{\rm o}_N(1)}$. This gives the same order of fluctuation as for the eigenvalues of matrices from the Gaussian Orthogonal Ensemble.Open AccessOptimal Eigenvalue Rigidity of Random Regular Graphs
Consider the normalized adjacency matrices of random $d$-regular graphs on $N$ vertices with fixed degree $d\geq 3$, and denote the eigenvalues as $λ_1=d/\sqrt{d-1}\geq λ_2\geqλ_3\cdots\geq λ_N$. We prove that the optimal (up to an extra $N^{{\rm o}_N(1)}$ factor, where ${\rm o}_N(1)$ can be arbitrarily small) eigenvalue rigidity holds. More precisely, denote $γ_i$ as the classical location of the $i$-th eigenvalue under the Kesten-Mckay law in decreasing order. Then with probability $1-N^{-1+{\rm o}_N(1)}$, \begin{align*} |λ_i-γ_i|\leq \frac{N^{{\rm o}_N(1)}}{N^{2/3} (\min\{i,N-i+1\})^{1/3}},\quad \text{ for all } i\in \{2,3,\cdots,N\}. \end{align*} In particular, the fluctuations of extreme eigenvalues are bounded by $N^{-2/3+{\rm o}_N(1)}$. This gives the same order of fluctuation as for the eigenvalues of matrices from the Gaussian Orthogonal Ensemble. - 14 February 2024
- Published by Oxford University Press (OUP) in International Mathematics Research Notices
- Vol. 2024 (7) , 6224-6258
Nodal Decompositions of a Symmetric Matrix
Analyzing nodal domains is a way to discern the structure of eigenvectors of operators on a graph. We give a new definition extending the concept of nodal domains to arbitrary signed graphs, and therefore to arbitrary symmetric matrices. We show that for an arbitrary symmetric matrix, a positive fraction of eigenbases satisfy a generalized version of known nodal bounds for un-signed (that is classical) graphs. We do this through an explicit decomposition. Moreover, we show that with high probability, the number of nodal domains of a bulk eigenvector of the adjacency matrix of a signed Erdős-Rényi graph is |$\Omega (n/\log n)$| and |$o(n)$|.
- 14 February 2024
- Published by Oxford University Press (OUP) in International Mathematics Research Notices
- Vol. 2024 (7) , 6224-6258
Nodal Decompositions of a Symmetric Matrix
Analyzing nodal domains is a way to discern the structure of eigenvectors of operators on a graph. We give a new definition extending the concept of nodal domains to arbitrary signed graphs, and therefore to arbitrary symmetric matrices. We show that for an arbitrary symmetric matrix, a positive fraction of eigenbases satisfy a generalized version of known nodal bounds for un-signed (that is classical) graphs. We do this through an explicit decomposition. Moreover, we show that with high probability, the number of nodal domains of a bulk eigenvector of the adjacency matrix of a signed Erdős-Rényi graph is |$\Omega (n/\log n)$| and |$o(n)$|.
- Open Access
Explicit two-sided unique-neighbor expanders
We study the problem of constructing explicit sparse graphs that exhibit strong vertex expansion. Our main result is the first two-sided construction of imbalanced unique-neighbor expanders, meaning bipartite graphs where small sets contained in both the left and right bipartitions exhibit unique-neighbor expansion, along with algebraic properties relevant to constructing quantum codes. Our constructions are obtained from instantiations of the tripartite line product of a large tripartite spectral expander and a sufficiently good constant-sized unique-neighbor expander, a new graph product we defined that generalizes the line product in the work of Alon and Capalbo and the routed product in the work of Asherov and Dinur. To analyze the vertex expansion of graphs arising from the tripartite line product, we develop a sharp characterization of subgraphs that can arise in bipartite spectral expanders, generalizing results of Kahale, which may be of independent interest. By picking appropriate graphs to apply our product to, we give a strongly explicit construction of an infinite family of $(d_1,d_2)$-biregular graphs $(G_n)_{n\ge 1}$ (for large enough $d_1$ and $d_2$) where all sets $S$ with fewer than a small constant fraction of vertices have $Ω(d_1\cdot |S|)$ unique-neighbors (assuming $d_1 \leq d_2$). Additionally, we can also guarantee that subsets of vertices of size up to $\exp(Ω(\sqrt{\log |V(G_n)|}))$ expand losslessly.Open AccessExplicit two-sided unique-neighbor expanders
We study the problem of constructing explicit sparse graphs that exhibit strong vertex expansion. Our main result is the first two-sided construction of imbalanced unique-neighbor expanders, meaning bipartite graphs where small sets contained in both the left and right bipartitions exhibit unique-neighbor expansion, along with algebraic properties relevant to constructing quantum codes. Our constructions are obtained from instantiations of the tripartite line product of a large tripartite spectral expander and a sufficiently good constant-sized unique-neighbor expander, a new graph product we defined that generalizes the line product in the work of Alon and Capalbo and the routed product in the work of Asherov and Dinur. To analyze the vertex expansion of graphs arising from the tripartite line product, we develop a sharp characterization of subgraphs that can arise in bipartite spectral expanders, generalizing results of Kahale, which may be of independent interest. By picking appropriate graphs to apply our product to, we give a strongly explicit construction of an infinite family of $(d_1,d_2)$-biregular graphs $(G_n)_{n\ge 1}$ (for large enough $d_1$ and $d_2$) where all sets $S$ with fewer than a small constant fraction of vertices have $Ω(d_1\cdot |S|)$ unique-neighbors (assuming $d_1 \leq d_2$). Additionally, we can also guarantee that subsets of vertices of size up to $\exp(Ω(\sqrt{\log |V(G_n)|}))$ expand losslessly. - Open Access
Nodal decompositions of a symmetric matrix
Analyzing nodal domains is a way to discern the structure of eigenvectors of operators on a graph. We give a new definition extending the concept of nodal domains to arbitrary signed graphs, and therefore to arbitrary symmetric matrices. We show that for an arbitrary symmetric matrix, a positive fraction of eigenbases satisfy a generalized version of known nodal bounds for un-signed (that is classical) graphs. We do this through an explicit decomposition. Moreover, we show that with high probability, the number of nodal domains of a bulk eigenvector of the adjacency matrix of signed a Erdős-Rényi graph is $Ω(n/\log n)$ and $o(n)$.Open AccessNodal decompositions of a symmetric matrix
Analyzing nodal domains is a way to discern the structure of eigenvectors of operators on a graph. We give a new definition extending the concept of nodal domains to arbitrary signed graphs, and therefore to arbitrary symmetric matrices. We show that for an arbitrary symmetric matrix, a positive fraction of eigenbases satisfy a generalized version of known nodal bounds for un-signed (that is classical) graphs. We do this through an explicit decomposition. Moreover, we show that with high probability, the number of nodal domains of a bulk eigenvector of the adjacency matrix of signed a Erdős-Rényi graph is $Ω(n/\log n)$ and $o(n)$. - Open Access
The Spectral Edge of Constant Degree Erdős-Rényi Graphs
We show that for an Erdős-Rényi graph on $N$ vertices with expected degree $d$ satisfying $\log^{-1/9}N\leq d\leq \log^{1/40}N$, the largest eigenvalues can be precisely determined by small neighborhoods around vertices of close to maximal degree. Moreover, under the added condition that $d\geq\log^{-1/15}N$, the corresponding eigenvectors are localized, in that the mass of the eigenvector decays exponentially away from the high degree vertex. This dependence on local neighborhoods implies that the edge eigenvalues converge to a Poisson point process. These theorems extend a result of Alt, Ducatez, and Knowles, who showed the same behavior for $d$ satisfying $(\log\log N)^4\ll d\leq (1-o_{N}(1))\frac{1}{\log 4-1}\log N$. To achieve high accuracy in the constant degree regime, instead of attempting to guess an approximate eigenvector of a local neighborhood, we analyze the true eigenvector of a local neighborhood, and show it must be localized and depend on local geometry.Open AccessThe Spectral Edge of Constant Degree Erdős-Rényi Graphs
We show that for an Erdős-Rényi graph on $N$ vertices with expected degree $d$ satisfying $\log^{-1/9}N\leq d\leq \log^{1/40}N$, the largest eigenvalues can be precisely determined by small neighborhoods around vertices of close to maximal degree. Moreover, under the added condition that $d\geq\log^{-1/15}N$, the corresponding eigenvectors are localized, in that the mass of the eigenvector decays exponentially away from the high degree vertex. This dependence on local neighborhoods implies that the edge eigenvalues converge to a Poisson point process. These theorems extend a result of Alt, Ducatez, and Knowles, who showed the same behavior for $d$ satisfying $(\log\log N)^4\ll d\leq (1-o_{N}(1))\frac{1}{\log 4-1}\log N$. To achieve high accuracy in the constant degree regime, instead of attempting to guess an approximate eigenvector of a local neighborhood, we analyze the true eigenvector of a local neighborhood, and show it must be localized and depend on local geometry. - 19 September 2023
- Published by Springer Nature in Communications in Mathematical Physics
- Vol. 403 (3) , 1477-1509
Quantum Ergodicity for Periodic Graphs
- 19 September 2023
- Published by Springer Nature in Communications in Mathematical Physics
- Vol. 403 (3) , 1477-1509
Quantum Ergodicity for Periodic Graphs
- Open Access
- 15 June 2021
- Published by Association for Computing Machinery (ACM)
- p. 396-407
Support of closed walks and second eigenvalue multiplicity of graphs
We show that the multiplicity of the second normalized adjacency matrix eigenvalue of any connected graph of maximum degree Δ is bounded by O(n Δ7/5/log1/5−o(1)n) for any Δ, and improve this to O(nlog1/2d/log1/4−o(1)n) for simple d-regular graphs when d≥ log1/4n. In fact, the same bounds hold for the number of eigenvalues in any interval of width λ2/logΔ1−o(1)n containing the second eigenvalue λ2. The main ingredient in the proof is a polynomial (in k) lower bound on the typical support of a closed random walk of length 2k in any connected graph, which in turn relies on new lower bounds for the entries of the Perron eigenvector of submatrices of the normalized adjacency matrix.
- 15 June 2021
- Published by Association for Computing Machinery (ACM)
- p. 396-407
Open AccessSupport of closed walks and second eigenvalue multiplicity of graphs
We show that the multiplicity of the second normalized adjacency matrix eigenvalue of any connected graph of maximum degree Δ is bounded by O(n Δ7/5/log1/5−o(1)n) for any Δ, and improve this to O(nlog1/2d/log1/4−o(1)n) for simple d-regular graphs when d≥ log1/4n. In fact, the same bounds hold for the number of eigenvalues in any interval of width λ2/logΔ1−o(1)n containing the second eigenvalue λ2. The main ingredient in the proof is a polynomial (in k) lower bound on the typical support of a closed random walk of length 2k in any connected graph, which in turn relies on new lower bounds for the entries of the Perron eigenvector of submatrices of the normalized adjacency matrix.