-
Fast Proxy Experiment Design for Causal Effect Identification
Authors:
Sepehr Elahi,
Sina Akbari,
Jalal Etesami,
Negar Kiyavash,
Patrick Thiran
Abstract:
Identifying causal effects is a key problem of interest across many disciplines. The two long-standing approaches to estimate causal effects are observational and experimental (randomized) studies. Observational studies can suffer from unmeasured confounding, which may render the causal effects unidentifiable. On the other hand, direct experiments on the target variable may be too costly or even i…
▽ More
Identifying causal effects is a key problem of interest across many disciplines. The two long-standing approaches to estimate causal effects are observational and experimental (randomized) studies. Observational studies can suffer from unmeasured confounding, which may render the causal effects unidentifiable. On the other hand, direct experiments on the target variable may be too costly or even infeasible to conduct. A middle ground between these two approaches is to estimate the causal effect of interest through proxy experiments, which are conducted on variables with a lower cost to intervene on compared to the main target. Akbari et al. [2022] studied this setting and demonstrated that the problem of designing the optimal (minimum-cost) experiment for causal effect identification is NP-complete and provided a naive algorithm that may require solving exponentially many NP-hard problems as a sub-routine in the worst case. In this work, we provide a few reformulations of the problem that allow for designing significantly more efficient algorithms to solve it as witnessed by our extensive simulations. Additionally, we study the closely-related problem of designing experiments that enable us to identify a given effect through valid adjustments sets.
△ Less
Submitted 7 July, 2024;
originally announced July 2024.
-
Spectral Methods for Matrix Product Factorization
Authors:
Saieed Akbari,
Yi-Zheng Fan,
Fu-Tao Hu,
Babak Miraftab,
Yi Wang
Abstract:
A graph $G$ is factored into graphs $H$ and $K$ via a matrix product if there exist adjacency matrices $A$, $B$, and $C$ of $G$, $H$, and $K$, respectively, such that $A = BC$. In this paper, we study the spectral aspects of the matrix product of graphs, including regularity, bipartiteness, and connectivity. We show that if a graph $G$ is factored into a connected graph $H$ and a graph $K$ with no…
▽ More
A graph $G$ is factored into graphs $H$ and $K$ via a matrix product if there exist adjacency matrices $A$, $B$, and $C$ of $G$, $H$, and $K$, respectively, such that $A = BC$. In this paper, we study the spectral aspects of the matrix product of graphs, including regularity, bipartiteness, and connectivity. We show that if a graph $G$ is factored into a connected graph $H$ and a graph $K$ with no isolated vertices, then certain properties hold. If $H$ is non-bipartite, then $G$ is connected. If $H$ is bipartite and $G$ is not connected, then $K$ is a regular bipartite graph, and consequently, $n$ is even. Furthermore, we show that trees are not factorizable, which answers a question posed by Maghsoudi et al.
△ Less
Submitted 4 July, 2024;
originally announced July 2024.
-
Line graphs and Nordhaus-Gaddum-type bounds for self-loop graphs
Authors:
Saieed Akbari,
Irena M. Jovanović,
Johnny Lim
Abstract:
Let $G_S$ be the graph obtained by attaching a self-loop at every vertex in $S \subseteq V(G)$ of a simple graph $G$ of order $n.$ In this paper, we explore several new results related to the line graph $L(G_S)$ of $G_S.$ Particularly, we show that every eigenvalue of $L(G_S)$ must be at least $-2,$ and relate the characteristic polynomial of the line graph $L(G)$ of $G$ with the characteristic po…
▽ More
Let $G_S$ be the graph obtained by attaching a self-loop at every vertex in $S \subseteq V(G)$ of a simple graph $G$ of order $n.$ In this paper, we explore several new results related to the line graph $L(G_S)$ of $G_S.$ Particularly, we show that every eigenvalue of $L(G_S)$ must be at least $-2,$ and relate the characteristic polynomial of the line graph $L(G)$ of $G$ with the characteristic polynomial of the line graph $L(\widehat{G})$ of a self-loop graph $\widehat{G}$, which is obtained by attaching a self-loop at each vertex of $G$. Then, we provide some new bounds for the eigenvalues and energy of $G_S.$ As one of the consequences, we obtain that the energy of a connected regular complete multipartite graph is not greater than the energy of the corresponding self-loop graph. Lastly, we establish a lower bound of the spectral radius in terms of the first Zagreb index $M_1(G)$ and the minimum degree $δ(G),$ as well as proving two Nordhaus-Gaddum-type bounds for the spectral radius and the energy of $G_S,$ respectively.
△ Less
Submitted 15 May, 2024;
originally announced May 2024.
-
Recursive Causal Discovery
Authors:
Ehsan Mokhtarian,
Sepehr Elahi,
Sina Akbari,
Negar Kiyavash
Abstract:
Causal discovery, i.e., learning the causal graph from data, is often the first step toward the identification and estimation of causal effects, a key requirement in numerous scientific domains. Causal discovery is hampered by two main challenges: limited data results in errors in statistical testing and the computational complexity of the learning task is daunting. This paper builds upon and exte…
▽ More
Causal discovery, i.e., learning the causal graph from data, is often the first step toward the identification and estimation of causal effects, a key requirement in numerous scientific domains. Causal discovery is hampered by two main challenges: limited data results in errors in statistical testing and the computational complexity of the learning task is daunting. This paper builds upon and extends four of our prior publications (Mokhtarian et al., 2021; Akbari et al., 2021; Mokhtarian et al., 2022, 2023a). These works introduced the concept of removable variables, which are the only variables that can be removed recursively for the purpose of causal discovery. Presence and identification of removable variables allow recursive approaches for causal discovery, a promising solution that helps to address the aforementioned challenges by reducing the problem size successively. This reduction not only minimizes conditioning sets in each conditional independence (CI) test, leading to fewer errors but also significantly decreases the number of required CI tests. The worst-case performances of these methods nearly match the lower bound. In this paper, we present a unified framework for the proposed algorithms, refined with additional details and enhancements for a coherent presentation. A comprehensive literature review is also included, comparing the computational complexity of our methods with existing approaches, showcasing their state-of-the-art efficiency. Another contribution of this paper is the release of RCD, a Python package that efficiently implements these algorithms. This package is designed for practitioners and researchers interested in applying these methods in practical scenarios. The package is available at github.com/ban-epfl/rcd, with comprehensive documentation provided at rcdpackage.com.
△ Less
Submitted 14 March, 2024;
originally announced March 2024.
-
On the $Δ$-edge stability number of graphs
Authors:
Saieed Akbari,
Reza Hosseini Dolatabadi,
Mohsen Jamaali,
Sandi Klavžar,
Nazanin Movarraei
Abstract:
The $Δ$-edge stability number ${\rm es}_Δ(G)$ of a graph $G$ is the minimum number of edges of $G$ whose removal results in a subgraph $H$ with $Δ(H) = Δ(G)-1$. Sets whose removal results in a subgraph with smaller maximum degree are called mitigating sets. It is proved that there always exists a mitigating set which induces a disjoint union of paths of order $2$ or $3$. Minimum mitigating sets wh…
▽ More
The $Δ$-edge stability number ${\rm es}_Δ(G)$ of a graph $G$ is the minimum number of edges of $G$ whose removal results in a subgraph $H$ with $Δ(H) = Δ(G)-1$. Sets whose removal results in a subgraph with smaller maximum degree are called mitigating sets. It is proved that there always exists a mitigating set which induces a disjoint union of paths of order $2$ or $3$. Minimum mitigating sets which induce matchings are characterized. It is proved that to obtain an upper bound of the form ${\rm es}_Δ(G) \leq c |V(G)|$ for an arbitrary graph $G$ of given maximum degree $Δ$, where $c$ is a given constant, it suffices to prove the bound for $Δ$-regular graphs. Sharp upper bounds of this form are derived for regular graphs. It is proved that if $Δ(G) \geq\frac{|V(G)|-2}{3}$ or the induced subgraph on maximum degree vertices has a $Δ(G)$-edge coloring, then ${\rm es}_Δ(G) \le |V(G)|/2$.
△ Less
Submitted 8 March, 2024;
originally announced March 2024.
-
Lower bounds for the Randić index in terms of matching number
Authors:
Saieed Akbari,
Sina Ghasemi Nezhad,
Reyhane Ghazizadeh,
John Haslegrave,
Elahe Tohidi
Abstract:
We investigate how small the Randić index of a graph can be in terms of its matching number, and prove several results. We give best-possible linear bounds for graphs of small excess and for subcubic graphs; in the former case the size of excess we permit is qualitatively the best possible. We show that a linear bound holds for any sparse hereditary graph class (such as planar graphs). In general,…
▽ More
We investigate how small the Randić index of a graph can be in terms of its matching number, and prove several results. We give best-possible linear bounds for graphs of small excess and for subcubic graphs; in the former case the size of excess we permit is qualitatively the best possible. We show that a linear bound holds for any sparse hereditary graph class (such as planar graphs). In general, however, we show that it can be much smaller than linear. We determine the asymptotic growth rate of the minimum Randić index for graphs with a near perfect matching, and conjecture that the same bounds hold for all graphs.
△ Less
Submitted 20 February, 2024;
originally announced February 2024.
-
Non-linear Triple Changes Estimator for Targeted Policies
Authors:
Sina Akbari,
Negar Kiyavash
Abstract:
The renowned difference-in-differences (DiD) estimator relies on the assumption of 'parallel trends,' which does not hold in many practical applications. To address this issue, the econometrics literature has turned to the triple difference estimator. Both DiD and triple difference are limited to assessing average effects exclusively. An alternative avenue is offered by the changes-in-changes (CiC…
▽ More
The renowned difference-in-differences (DiD) estimator relies on the assumption of 'parallel trends,' which does not hold in many practical applications. To address this issue, the econometrics literature has turned to the triple difference estimator. Both DiD and triple difference are limited to assessing average effects exclusively. An alternative avenue is offered by the changes-in-changes (CiC) estimator, which provides an estimate of the entire counterfactual distribution at the cost of relying on (stronger) distributional assumptions. In this work, we extend the triple difference estimator to accommodate the CiC framework, presenting the `triple changes estimator' and its identification assumptions, thereby expanding the scope of the CiC paradigm. Subsequently, we empirically evaluate the proposed framework and apply it to a study examining the impact of Medicaid expansion on children's preventive care.
△ Less
Submitted 19 February, 2024;
originally announced February 2024.
-
Lyapunov-Based Dropout Deep Neural Network (Lb-DDNN) Controller
Authors:
Saiedeh Akbari,
Emily J. Griffis,
Omkar Sudhir Patil,
Warren E. Dixon
Abstract:
Deep neural network (DNN)-based adaptive controllers can be used to compensate for unstructured uncertainties in nonlinear dynamic systems. However, DNNs are also very susceptible to overfitting and co-adaptation. Dropout regularization is an approach where nodes are randomly dropped during training to alleviate issues such as overfitting and co-adaptation. In this paper, a dropout DNN-based adapt…
▽ More
Deep neural network (DNN)-based adaptive controllers can be used to compensate for unstructured uncertainties in nonlinear dynamic systems. However, DNNs are also very susceptible to overfitting and co-adaptation. Dropout regularization is an approach where nodes are randomly dropped during training to alleviate issues such as overfitting and co-adaptation. In this paper, a dropout DNN-based adaptive controller is developed. The developed dropout technique allows the deactivation of weights that are stochastically selected for each individual layer within the DNN. Simultaneously, a Lyapunov-based real-time weight adaptation law is introduced to update the weights of all layers of the DNN for online unsupervised learning. A non-smooth Lyapunov-based stability analysis is performed to ensure asymptotic convergence of the tracking error. Simulation results of the developed dropout DNN-based adaptive controller indicate a 38.32% improvement in the tracking error, a 53.67% improvement in the function approximation error, and 50.44% lower control effort when compared to a baseline adaptive DNN-based controller without dropout regularization.
△ Less
Submitted 30 October, 2023;
originally announced October 2023.
-
2-Coupon Coloring of Cubic Graphs Containing 3-Cycle or 4-Cycle
Authors:
S. Akbari,
M. Azimian,
A. Fazli Khani,
B. Samimi,
E. Zahiri
Abstract:
Let $G$ be a graph. A total dominating set in a graph $G$ is a set $S$ of vertices of $G$ such that every vertex in $G$ is adjacent to a vertex in $S$. Recently, the following question was proposed: "Is it true that every connected cubic graph containing a $3$-cycle has two vertex disjoint total dominating sets?" In this paper, we give a negative answer to this question. Moreover, we prove that if…
▽ More
Let $G$ be a graph. A total dominating set in a graph $G$ is a set $S$ of vertices of $G$ such that every vertex in $G$ is adjacent to a vertex in $S$. Recently, the following question was proposed: "Is it true that every connected cubic graph containing a $3$-cycle has two vertex disjoint total dominating sets?" In this paper, we give a negative answer to this question. Moreover, we prove that if we replace $3$-cycle with $4$-cycle the answer is affirmative. This implies every connected cubic graph containing a diamond (the complete graph of order $4$ minus one edge) as a subgraph can be partitioned into two total dominating sets, a result that was proved in 2017.
△ Less
Submitted 29 August, 2023;
originally announced August 2023.
-
Causal Imitability Under Context-Specific Independence Relations
Authors:
Fateme Jamshidi,
Sina Akbari,
Negar Kiyavash
Abstract:
Drawbacks of ignoring the causal mechanisms when performing imitation learning have recently been acknowledged. Several approaches both to assess the feasibility of imitation and to circumvent causal confounding and causal misspecifications have been proposed in the literature. However, the potential benefits of the incorporation of additional information about the underlying causal structure are…
▽ More
Drawbacks of ignoring the causal mechanisms when performing imitation learning have recently been acknowledged. Several approaches both to assess the feasibility of imitation and to circumvent causal confounding and causal misspecifications have been proposed in the literature. However, the potential benefits of the incorporation of additional information about the underlying causal structure are left unexplored. An example of such overlooked information is context-specific independence (CSI), i.e., independence that holds only in certain contexts. We consider the problem of causal imitation learning when CSI relations are known. We prove that the decision problem pertaining to the feasibility of imitation in this setting is NP-hard. Further, we provide a necessary graphical criterion for imitation learning under CSI and show that under a structural assumption, this criterion is also sufficient. Finally, we propose a sound algorithmic approach for causal imitation learning which takes both CSI relations and data into account.
△ Less
Submitted 11 June, 2023; v1 submitted 1 June, 2023;
originally announced June 2023.
-
Learning Causal Graphs via Monotone Triangular Transport Maps
Authors:
Sina Akbari,
Luca Ganassali,
Negar Kiyavash
Abstract:
We study the problem of causal structure learning from data using optimal transport (OT). Specifically, we first provide a constraint-based method which builds upon lower-triangular monotone parametric transport maps to design conditional independence tests which are agnostic to the noise distribution. We provide an algorithm for causal discovery up to Markov Equivalence with no assumptions on the…
▽ More
We study the problem of causal structure learning from data using optimal transport (OT). Specifically, we first provide a constraint-based method which builds upon lower-triangular monotone parametric transport maps to design conditional independence tests which are agnostic to the noise distribution. We provide an algorithm for causal discovery up to Markov Equivalence with no assumptions on the structural equations/noise distributions, which allows for settings with latent variables. Our approach also extends to score-based causal discovery by providing a novel means for defining scores. This allows us to uniquely recover the causal graph under additional identifiability and structural assumptions, such as additive noise or post-nonlinear models. We provide experimental results to compare the proposed approach with the state of the art on both synthetic and real-world datasets.
△ Less
Submitted 26 May, 2023;
originally announced May 2023.
-
Some Results On Spectrum And Energy Of Graphs With Loops
Authors:
Saieed Akbari,
Hussah Al Menderj,
Miin Huey Ang,
Johnny Lim,
Zhen Chuan Ng
Abstract:
Let $G_S$ be a graph with loops obtained from a graph $G$ of order $n$ and loops at $S \subseteq V(G)$. In this paper, we establish a neccesary and sufficient condition on the bipartititeness of a connected graph $G$ and the spectrum Spec($G_S$) and Spec($G_{V(G)\backslash S}$). We also prove that for every $S \subseteq V(G)$, $E(G_S) \geq E(G)$ when $G$ is bipartite. Moreover, we provide an ident…
▽ More
Let $G_S$ be a graph with loops obtained from a graph $G$ of order $n$ and loops at $S \subseteq V(G)$. In this paper, we establish a neccesary and sufficient condition on the bipartititeness of a connected graph $G$ and the spectrum Spec($G_S$) and Spec($G_{V(G)\backslash S}$). We also prove that for every $S \subseteq V(G)$, $E(G_S) \geq E(G)$ when $G$ is bipartite. Moreover, we provide an identification of the spectrum of complete graphs $K_n$ and complete bipartite graphs $K_{m,n}$ with loops. We characterize any graphs with loops of order n whose eigenvalues are all positive or non-negative, and also any graphs with a few distinct eigenvalues. Finally, we provide some bounds related to $G_S$.
△ Less
Submitted 11 April, 2023;
originally announced April 2023.
-
Numerical assessments of a nonintrusive surrogate model based on recurrent neural networks and proper orthogonal decomposition: Rayleigh Benard convection
Authors:
Saeed Akbari,
Suraj Pawar,
Omer San
Abstract:
Recent developments in diagnostic and computing technologies offer to leverage numerous forms of nonintrusive modeling approaches from data where machine learning can be used to build computationally cheap and accurate surrogate models. To this end, we present a nonlinear proper orthogonal decomposition (POD) framework, denoted as NLPOD, to forge a nonintrusive reduced-order model for the Boussine…
▽ More
Recent developments in diagnostic and computing technologies offer to leverage numerous forms of nonintrusive modeling approaches from data where machine learning can be used to build computationally cheap and accurate surrogate models. To this end, we present a nonlinear proper orthogonal decomposition (POD) framework, denoted as NLPOD, to forge a nonintrusive reduced-order model for the Boussinesq equations. In our NLPOD approach, we first employ the POD procedure to obtain a set of global modes to build a linear-fit latent space and utilize an autoencoder network to compress the projection of this latent space through a nonlinear unsupervised map** of POD coefficients. Then, long short-term memory (LSTM) neural network architecture is utilized to discover temporal patterns in this low-rank manifold. While performing a detailed sensitivity analysis for hyperparameters of the LSTM model, the trade-off between accuracy and efficiency is systematically analyzed for solving a canonical Rayleigh-Benard convection system.
△ Less
Submitted 10 December, 2022;
originally announced December 2022.
-
End Super Dominating Sets in Graphs
Authors:
Saieed Akbari,
Nima Ghanbari,
Michael A. Henning
Abstract:
Let $G=(V,E)$ be a simple graph. A dominating set of $G$ is a subset $S\subseteq V$ such that every vertex not in $S$ is adjacent to at least one vertex in $S$. The cardinality of a smallest dominating set of $G$, denoted by $γ(G)$, is the domination number of $G$. Two vertices are neighbors if they are adjacent. A super dominating set is a dominating set $S$ with the additional property that ever…
▽ More
Let $G=(V,E)$ be a simple graph. A dominating set of $G$ is a subset $S\subseteq V$ such that every vertex not in $S$ is adjacent to at least one vertex in $S$. The cardinality of a smallest dominating set of $G$, denoted by $γ(G)$, is the domination number of $G$. Two vertices are neighbors if they are adjacent. A super dominating set is a dominating set $S$ with the additional property that every vertex in $V \setminus S$ has a neighbor in $S$ that is adjacent to no other vertex in $V \setminus S$. Moreover if every vertex in $V \setminus S$ has degree at least~$2$, then $S$ is an end super dominating set. The end super domination number is the minimum cardinality of an end super dominating set. We give applications of end super dominating sets as main servers and temporary servers of networks. We determine the exact value of the end super domination number for specific classes of graphs, and we count the number of end super dominating sets in these graphs. Tight upper bounds on the end super domination number are established, where the graph is modified by vertex (edge) removal and contraction.
△ Less
Submitted 14 November, 2022; v1 submitted 7 September, 2022;
originally announced September 2022.
-
Causal Effect Identification in Uncertain Causal Networks
Authors:
Sina Akbari,
Fateme Jamshidi,
Ehsan Mokhtarian,
Matthew J. Vowels,
Jalal Etesami,
Negar Kiyavash
Abstract:
Causal identification is at the core of the causal inference literature, where complete algorithms have been proposed to identify causal queries of interest. The validity of these algorithms hinges on the restrictive assumption of having access to a correctly specified causal structure. In this work, we study the setting where a probabilistic model of the causal structure is available. Specificall…
▽ More
Causal identification is at the core of the causal inference literature, where complete algorithms have been proposed to identify causal queries of interest. The validity of these algorithms hinges on the restrictive assumption of having access to a correctly specified causal structure. In this work, we study the setting where a probabilistic model of the causal structure is available. Specifically, the edges in a causal graph exist with uncertainties which may, for example, represent degree of belief from domain experts. Alternatively, the uncertainty about an edge may reflect the confidence of a particular statistical test. The question that naturally arises in this setting is: Given such a probabilistic graph and a specific causal effect of interest, what is the subgraph which has the highest plausibility and for which the causal effect is identifiable? We show that answering this question reduces to solving an NP-complete combinatorial optimization problem which we call the edge ID problem. We propose efficient algorithms to approximate this problem and evaluate them against both real-world networks and randomly generated graphs.
△ Less
Submitted 27 October, 2023; v1 submitted 9 August, 2022;
originally announced August 2022.
-
A lower bound of the energy of non-singular graphs in terms of average degree
Authors:
Saieed Akbari,
Hossein Dabirian,
S. Mahmood Ghasemi
Abstract:
Let $G$ be a graph of order $n$ with adjacency matrix $A(G)$. The \textit{energy} of graph $G$, denoted by $\mathcal{E}(G)$, is defined as the sum of absolute value of eigenvalues of $A(G)$. It was conjectured that if $A(G)$ is non-singular, then $\mathcal{E}(G)\geqΔ(G)+δ(G)$. In this paper we propose a stronger conjecture as for $n \geq 5$, $\mathcal{E}(G)\geq n-1+ d$, where $d$ is the average de…
▽ More
Let $G$ be a graph of order $n$ with adjacency matrix $A(G)$. The \textit{energy} of graph $G$, denoted by $\mathcal{E}(G)$, is defined as the sum of absolute value of eigenvalues of $A(G)$. It was conjectured that if $A(G)$ is non-singular, then $\mathcal{E}(G)\geqΔ(G)+δ(G)$. In this paper we propose a stronger conjecture as for $n \geq 5$, $\mathcal{E}(G)\geq n-1+ d$, where $d$ is the average degree of $G$. Here, we show that conjecture holds for bipartite graphs, planar graphs and for the graphs with $d \leq n-2\ln n -3$
△ Less
Submitted 10 July, 2022;
originally announced July 2022.
-
Tight Bounds on the Chromatic Edge Stability Index of Graphs
Authors:
Saieed Akbari,
John Haslegrave,
Mehrbod Javadi,
Nasim Nahvi,
Helia Niaparast
Abstract:
The chromatic edge stability index $\mathrm{es}_{χ'}(G)$ of a graph $G$ is the minimum number of edges whose removal results in a graph with smaller chromatic index. We give best-possible upper bounds on $\mathrm{es}_{χ'}(G)$ in terms of the number of vertices of degree $Δ(G)$ (if $G$ is Class 2), and the numbers of vertices of degree $Δ(G)$ and ${Δ(G)-1}$ (if $G$ is Class 1). If $G$ is bipartite…
▽ More
The chromatic edge stability index $\mathrm{es}_{χ'}(G)$ of a graph $G$ is the minimum number of edges whose removal results in a graph with smaller chromatic index. We give best-possible upper bounds on $\mathrm{es}_{χ'}(G)$ in terms of the number of vertices of degree $Δ(G)$ (if $G$ is Class 2), and the numbers of vertices of degree $Δ(G)$ and ${Δ(G)-1}$ (if $G$ is Class 1). If $G$ is bipartite we give an exact expression for $\mathrm{es}_{χ'}(G)$ involving the maximum size of a matching in the subgraph induced by vertices of degree $Δ(G)$. Finally, we consider whether a minimum mitigating set, that is a set of size $\mathrm{es}_{χ'}(G)$ whose removal reduces the chromatic index, has the property that every edge meets a vertex of degree at least $Δ(G)-1$; we prove that this is true for some minimum mitigating set of $G$, but not necessarily for every minimum mitigating set of $G$.
△ Less
Submitted 21 December, 2023; v1 submitted 8 June, 2022;
originally announced June 2022.
-
Experimental Design for Causal Effect Identification
Authors:
Sina Akbari,
Jalal Etesami,
Negar Kiyavash
Abstract:
Pearl's do calculus is a complete axiomatic approach to learn the identifiable causal effects from observational data. When such an effect is not identifiable, it is necessary to perform a collection of often costly interventions in the system to learn the causal effect. In this work, we consider the problem of designing the collection of interventions with the minimum cost to identify the desired…
▽ More
Pearl's do calculus is a complete axiomatic approach to learn the identifiable causal effects from observational data. When such an effect is not identifiable, it is necessary to perform a collection of often costly interventions in the system to learn the causal effect. In this work, we consider the problem of designing the collection of interventions with the minimum cost to identify the desired effect. First, we prove that this problem is NP-hard, and subsequently propose an algorithm that can either find the optimal solution or a logarithmic-factor approximation of it. This is done by establishing a connection between our problem and the minimum hitting set problem. Additionally, we propose several polynomial-time heuristic algorithms to tackle the computational complexity of the problem. Although these algorithms could potentially stumble on sub-optimal solutions, our simulations show that they achieve small regrets on random graphs.
△ Less
Submitted 17 August, 2023; v1 submitted 4 May, 2022;
originally announced May 2022.
-
A Free Lunch with Influence Functions? Improving Neural Network Estimates with Concepts from Semiparametric Statistics
Authors:
Matthew J. Vowels,
Sina Akbari,
Necati Cihan Camgoz,
Richard Bowden
Abstract:
Parameter estimation in empirical fields is usually undertaken using parametric models, and such models readily facilitate statistical inference. Unfortunately, they are unlikely to be sufficiently flexible to be able to adequately model real-world phenomena, and may yield biased estimates. Conversely, non-parametric approaches are flexible but do not readily facilitate statistical inference and m…
▽ More
Parameter estimation in empirical fields is usually undertaken using parametric models, and such models readily facilitate statistical inference. Unfortunately, they are unlikely to be sufficiently flexible to be able to adequately model real-world phenomena, and may yield biased estimates. Conversely, non-parametric approaches are flexible but do not readily facilitate statistical inference and may still exhibit residual bias. We explore the potential for Influence Functions (IFs) to (a) improve initial estimators without needing more data (b) increase model robustness and (c) facilitate statistical inference. We begin with a broad introduction to IFs, and propose a neural network method 'MultiNet', which seeks the diversity of an ensemble using a single architecture. We also introduce variants on the IF update step which we call 'MultiStep', and provide a comprehensive evaluation of different approaches. The improvements are found to be dataset dependent, indicating an interaction between the methods used and nature of the data generating process. Our experiments highlight the need for practitioners to check the consistency of their findings, potentially by undertaking multiple analyses with different combinations of estimators. We also show that it is possible to improve existing neural networks for `free', without needing more data, and without needing to retrain them.
△ Less
Submitted 10 June, 2022; v1 submitted 18 February, 2022;
originally announced February 2022.
-
Learning Bayesian Networks in the Presence of Structural Side Information
Authors:
Ehsan Mokhtarian,
Sina Akbari,
Fateme Jamshidi,
Jalal Etesami,
Negar Kiyavash
Abstract:
We study the problem of learning a Bayesian network (BN) of a set of variables when structural side information about the system is available. It is well known that learning the structure of a general BN is both computationally and statistically challenging. However, often in many applications, side information about the underlying structure can potentially reduce the learning complexity. In this…
▽ More
We study the problem of learning a Bayesian network (BN) of a set of variables when structural side information about the system is available. It is well known that learning the structure of a general BN is both computationally and statistically challenging. However, often in many applications, side information about the underlying structure can potentially reduce the learning complexity. In this paper, we develop a recursive constraint-based algorithm that efficiently incorporates such knowledge (i.e., side information) into the learning process. In particular, we study two types of structural side information about the underlying BN: (I) an upper bound on its clique number is known, or (II) it is diamond-free. We provide theoretical guarantees for the learning algorithms, including the worst-case number of tests required in each scenario. As a consequence of our work, we show that bounded treewidth BNs can be learned with polynomial complexity. Furthermore, we evaluate the performance and the scalability of our algorithms in both synthetic and real-world structures and show that they outperform the state-of-the-art structure learning algorithms.
△ Less
Submitted 20 December, 2021;
originally announced December 2021.
-
Recursive Causal Structure Learning in the Presence of Latent Variables and Selection Bias
Authors:
Sina Akbari,
Ehsan Mokhtarian,
AmirEmad Ghassami,
Negar Kiyavash
Abstract:
We consider the problem of learning the causal MAG of a system from observational data in the presence of latent variables and selection bias. Constraint-based methods are one of the main approaches for solving this problem, but the existing methods are either computationally impractical when dealing with large graphs or lacking completeness guarantees. We propose a novel computationally efficient…
▽ More
We consider the problem of learning the causal MAG of a system from observational data in the presence of latent variables and selection bias. Constraint-based methods are one of the main approaches for solving this problem, but the existing methods are either computationally impractical when dealing with large graphs or lacking completeness guarantees. We propose a novel computationally efficient recursive constraint-based method that is sound and complete. The key idea of our approach is that at each iteration a specific type of variable is identified and removed. This allows us to learn the structure efficiently and recursively, as this technique reduces both the number of required conditional independence (CI) tests and the size of the conditioning sets. The former substantially reduces the computational complexity, while the latter results in more reliable CI tests. We provide an upper bound on the number of required CI tests in the worst case. To the best of our knowledge, this is the tightest bound in the literature. We further provide a lower bound on the number of CI tests required by any constraint-based method. The upper bound of our proposed approach and the lower bound at most differ by a factor equal to the number of variables in the worst case. We provide experimental results to compare the proposed approach with the state of the art on both synthetic and real-world structures.
△ Less
Submitted 22 October, 2021;
originally announced October 2021.
-
A bound for the $p$-domination number of a graph in terms of its eigenvalue multiplicities
Authors:
A. Abiad,
S. Akbari,
M. H. Fakharan,
A. Mehdizadeh
Abstract:
Let $G$ be a connected graph of order $n$ with domination number $γ(G)$. Wang, Yan, Fang, Geng and Tian [Linear Algebra Appl. 607 (2020), 307-318] showed that for any Laplacian eigenvalue $λ$ of $G$ with multiplicity $m_G(λ)$, it holds that $γ(G)\leq n-m_G(λ)$. Using techniques from the theory of star sets, in this work we prove that the same bound holds when $λ$ is an arbitrary adjacency eigenval…
▽ More
Let $G$ be a connected graph of order $n$ with domination number $γ(G)$. Wang, Yan, Fang, Geng and Tian [Linear Algebra Appl. 607 (2020), 307-318] showed that for any Laplacian eigenvalue $λ$ of $G$ with multiplicity $m_G(λ)$, it holds that $γ(G)\leq n-m_G(λ)$. Using techniques from the theory of star sets, in this work we prove that the same bound holds when $λ$ is an arbitrary adjacency eigenvalue of a non-regular graph, and we characterize the cases of equality. Moreover, we show a result that gives a relationship between start sets and the $p$-domination number, and we apply it to extend the aforementioned spectral bound to the $p$-domination number using the adjacency and Laplacian eigenvalue multiplicities.
△ Less
Submitted 13 September, 2021;
originally announced September 2021.
-
On the Chromatic Vertex Stability Number of Graphs
Authors:
Saieed Akbari,
Arash Beikmohammadi,
Sandi Klavžar,
Nazanin Movarraei
Abstract:
The chromatic vertex (resp.\ edge) stability number ${\rm vs}_χ(G)$ (resp.\ ${\rm es}_χ(G)$) of a graph $G$ is the minimum number of vertices (resp.\ edges) whose deletion results in a graph $H$ with $χ(H)=χ(G)-1$. In the main result it is proved that if $G$ is a graph with $χ(G) \in \{ Δ(G), Δ(G)+1 \}$, then ${\rm vs}_χ(G) = {\rm ivs}_χ(G)$, where ${\rm ivs}_χ(G)$ is the independent chromatic ver…
▽ More
The chromatic vertex (resp.\ edge) stability number ${\rm vs}_χ(G)$ (resp.\ ${\rm es}_χ(G)$) of a graph $G$ is the minimum number of vertices (resp.\ edges) whose deletion results in a graph $H$ with $χ(H)=χ(G)-1$. In the main result it is proved that if $G$ is a graph with $χ(G) \in \{ Δ(G), Δ(G)+1 \}$, then ${\rm vs}_χ(G) = {\rm ivs}_χ(G)$, where ${\rm ivs}_χ(G)$ is the independent chromatic vertex stability number. The result need not hold for graphs $G$ with $χ(G) \le \frac{Δ(G)+1}{2}$. It is proved that if $χ(G) > \frac{Δ(G)}{2}+1$, then ${\rm vs}_χ(G) = {\rm es}_χ(G)$. A Nordhaus-Gaddum-type result on the chromatic vertex stability number is also given.
△ Less
Submitted 14 December, 2021; v1 submitted 30 August, 2021;
originally announced August 2021.
-
On the chromatic edge stability index of graphs
Authors:
Saieed Akbari,
Arash Beikmohammadi,
Boštjan Brešar,
Tanja Dravec,
Mohammad Mahdi Habibollahi,
Nazanin Movarraei
Abstract:
Given a non-trivial graph $G$, the minimum cardinality of a set of edges $F$ in $G$ such that $χ'(G \setminus F)<χ'(G)$ is called the chromatic edge stability index of $G$, denoted by $es_{χ'}(G)$, and such a (smallest) set $F$ is called a (minimum) mitigating set. While $1\le es_{χ'}(G)\le \lfloor n/2\rfloor$ holds for any graph $G$, we investigate the graphs with extremal and near-extremal value…
▽ More
Given a non-trivial graph $G$, the minimum cardinality of a set of edges $F$ in $G$ such that $χ'(G \setminus F)<χ'(G)$ is called the chromatic edge stability index of $G$, denoted by $es_{χ'}(G)$, and such a (smallest) set $F$ is called a (minimum) mitigating set. While $1\le es_{χ'}(G)\le \lfloor n/2\rfloor$ holds for any graph $G$, we investigate the graphs with extremal and near-extremal values of $es_{χ'}(G)$. The graphs $G$ with $es_{χ'}(G)=\lfloor n/2\rfloor$ are classified, and the graphs $G$ with $es_{χ'}(G)=\lfloor n/2\rfloor-1$ and $χ'(G)=Δ(G)+1$ are characterized. We establish that the odd cycles and $K_2$ are exactly the regular connected graphs with the chromatic edge stability index $1$; on the other hand, we prove that it is NP-hard to verify whether a graph $G$ has $es_{χ'}(G)=1$. We also prove that every minimum mitigating set of an $r$-regular graph $G$, where $r\ne 4$, with $es_{χ'}(G)=2$ is a matching. Furthermore, we propose a conjecture that for every graph $G$ there exists a minimum mitigating set, which is a matching, and prove that the conjecture holds for graphs $G$ with $es_{χ'}(G)\in\{1,2,\lfloor n/2\rfloor-1,\lfloor n/2\rfloor\}$, and for bipartite graphs.
△ Less
Submitted 24 August, 2021;
originally announced August 2021.
-
CAIM: Cooperative Angle of Arrival Estimation using the Ising Method
Authors:
Shiva Akbari,
Shahrokh Valaee
Abstract:
This paper proposes a cooperative angle-of-arrival(AoA) estimation, taking advantage of co-processing channel state information (CSI) from a group of access points that receive signals of the same source. Since received signals are sparse, we use Compressive Sensing (CS) to address the AoA estimation problem. We formulate this problem as a penalized l0-norm minimization, reformulate it as an Ising…
▽ More
This paper proposes a cooperative angle-of-arrival(AoA) estimation, taking advantage of co-processing channel state information (CSI) from a group of access points that receive signals of the same source. Since received signals are sparse, we use Compressive Sensing (CS) to address the AoA estimation problem. We formulate this problem as a penalized l0-norm minimization, reformulate it as an Ising energy problem, and solve it using Markov Chain Monte Carlo (MCMC). Simulation results show that our proposed method outperforms the existing methods in the literature.
△ Less
Submitted 27 April, 2021;
originally announced April 2021.
-
Spectra of strongly Deza graphs
Authors:
Saieed Akbari,
Willem H. Haemers,
Mohammad Ali Hosseinzadeh,
Vladislav V. Kabanov,
Elena V. Konstantinova,
Leonid Shalaginov
Abstract:
A Deza graph $G$ with parameters $(n,k,b,a)$ is a $k$-regular graph with $n$ vertices such that any two distinct vertices have $b$ or $a$ common neighbours. The children $G_A$ and $G_B$ of a Deza graph $G$ are defined on the vertex set of $G$ such that every two distinct vertices are adjacent in $G_A$ or $G_B$ if and only if they have $a$ or $b$ common neighbours, respectively. A strongly Deza gra…
▽ More
A Deza graph $G$ with parameters $(n,k,b,a)$ is a $k$-regular graph with $n$ vertices such that any two distinct vertices have $b$ or $a$ common neighbours. The children $G_A$ and $G_B$ of a Deza graph $G$ are defined on the vertex set of $G$ such that every two distinct vertices are adjacent in $G_A$ or $G_B$ if and only if they have $a$ or $b$ common neighbours, respectively. A strongly Deza graph is a Deza graph with strongly regular children. In this paper we give a spectral characterisation of strongly Deza graphs, show relationships between eigenvalues, and study strongly Deza graphs which are distance-regular.
△ Less
Submitted 18 January, 2021;
originally announced January 2021.
-
Zero-sum flows for Steiner systems
Authors:
Saieed Akbari,
Hamid Reza Maimani,
Leila Parsaei Majd,
Ian M. Wanless
Abstract:
Given a $t$-$(v, k, λ)$ design, $\mathcal{D}=(X,\mathcal{B})$, a zero-sum $n$-flow of $\mathcal{D}$ is a map $f : \mathcal{B}\longrightarrow \{\pm1,\ldots, \pm(n-1)\}$ such that for any point $x\in X$, the sum of $f$ over all blocks incident with $x$ is zero. For a positive integer $k$, we find a zero-sum $k$-flow for an STS$(u w)$ and for an STS$(2v+7)$ for $v\equiv 1~(\mathrm{mod}~4)$, if there…
▽ More
Given a $t$-$(v, k, λ)$ design, $\mathcal{D}=(X,\mathcal{B})$, a zero-sum $n$-flow of $\mathcal{D}$ is a map $f : \mathcal{B}\longrightarrow \{\pm1,\ldots, \pm(n-1)\}$ such that for any point $x\in X$, the sum of $f$ over all blocks incident with $x$ is zero. For a positive integer $k$, we find a zero-sum $k$-flow for an STS$(u w)$ and for an STS$(2v+7)$ for $v\equiv 1~(\mathrm{mod}~4)$, if there are STS$(u)$, STS$(w)$ and STS$(v)$ such that the STS$(u)$ and STS$(v)$ both have a zero-sum $k$-flow. In 2015, it was conjectured that for $v>7$ every STS$(v)$ admits a zero-sum $3$-flow. Here, it is shown that many cyclic STS$(v)$ have a zero-sum $3$-flow. Also, we investigate the existence of zero-sum flows for some Steiner quadruple systems.
△ Less
Submitted 4 January, 2021;
originally announced January 2021.
-
On a question of Haemers regarding vectors in the nullspace of Seidel matrices
Authors:
Saieed Akbari,
Sebastian M. Cioabă,
Samira Goudarzi,
Aidin Niaparast,
Artin Tajdini
Abstract:
In 2011, Haemers asked the following question: If $S$ is the Seidel matrix of a graph of order $n$ and $S$ is singular, does there exist an eigenvector of $S$ corresponding to $0$ which has only $\pm 1$ elements?
In this paper, we construct infinite families of graphs which give a negative answer to this question. One of our constructions implies that for every natural number $N$, there exists a…
▽ More
In 2011, Haemers asked the following question: If $S$ is the Seidel matrix of a graph of order $n$ and $S$ is singular, does there exist an eigenvector of $S$ corresponding to $0$ which has only $\pm 1$ elements?
In this paper, we construct infinite families of graphs which give a negative answer to this question. One of our constructions implies that for every natural number $N$, there exists a graph whose Seidel matrix $S$ is singular such that for any integer vector in the nullspace of $S$, the absolute value of any entry in this vector is more than $N$. We also derive some characteristics of vectors in the nullspace of Seidel matrices, which lead to some necessary conditions for the singularity of Seidel matrices. Finally, we obtain some properties of the graphs which affirm the above question.
△ Less
Submitted 21 January, 2021; v1 submitted 12 November, 2020;
originally announced November 2020.
-
A Recursive Markov Boundary-Based Approach to Causal Structure Learning
Authors:
Ehsan Mokhtarian,
Sina Akbari,
AmirEmad Ghassami,
Negar Kiyavash
Abstract:
Constraint-based methods are one of the main approaches for causal structure learning that are particularly valued as they are asymptotically guaranteed to find a structure that is Markov equivalent to the causal graph of the system. On the other hand, they may require an exponentially large number of conditional independence (CI) tests in the number of variables of the system. In this paper, we p…
▽ More
Constraint-based methods are one of the main approaches for causal structure learning that are particularly valued as they are asymptotically guaranteed to find a structure that is Markov equivalent to the causal graph of the system. On the other hand, they may require an exponentially large number of conditional independence (CI) tests in the number of variables of the system. In this paper, we propose a novel recursive constraint-based method for causal structure learning that significantly reduces the required number of CI tests compared to the existing literature. The idea of the proposed approach is to use Markov boundary information to identify a specific variable that can be removed from the set of variables without affecting the statistical dependencies among the other variables. Having identified such a variable, we discover its neighborhood, remove that variable from the set of variables, and recursively learn the causal structure over the remaining variables. We further provide a lower bound on the number of CI tests required by any constraint-based method. Comparing this lower bound to our achievable bound demonstrates the efficiency of the proposed approach. Our experimental results show that the proposed algorithm outperforms state-of-the-art both on synthetic and real-world structures.
△ Less
Submitted 20 May, 2021; v1 submitted 10 October, 2020;
originally announced October 2020.
-
Manufacture and Characterization of Graphene Membranes with Suspended Silicon Proof Masses for MEMS and NEMS Applications
Authors:
Xuge Fan,
Anderson D. Smith,
Fredrik Forsberg,
Stefan Wagner,
Stephan Schröder,
Sayedeh Shirin Afyouni Akbari,
Andreas C. Fischer,
Luis Guillermo Villanueva,
Mikael Östling,
Max C. Lemme,
Frank Niklaus
Abstract:
Unparalleled strength, chemical stability, ultimate surface-to-volume ratio and excellent electronic properties of graphene make it an ideal candidate as a material for membranes in micro- and nanoelectromechanical systems (MEMS and NEMS). However, the integration of graphene into MEMS or NEMS devices and suspended structures such as proof masses on graphene membranes raises several technological…
▽ More
Unparalleled strength, chemical stability, ultimate surface-to-volume ratio and excellent electronic properties of graphene make it an ideal candidate as a material for membranes in micro- and nanoelectromechanical systems (MEMS and NEMS). However, the integration of graphene into MEMS or NEMS devices and suspended structures such as proof masses on graphene membranes raises several technological challenges, including collapse and rupture of the graphene. We have developed a robust route for realizing membranes made of double-layer CVD graphene and suspending large silicon proof masses on membranes with high yields. We have demonstrated the manufacture of square graphene membranes with side lengths from 7 micro meter to 110 micro meter and suspended proof masses consisting of solid silicon cubes that are from 5 micro meter multiply 5 micro meter multiply 16.4 micro meter to 100 micro meter multiply 100 micro meter multiply 16.4 micro meter in size. Our approach is compatible with wafer-scale MEMS and semiconductor manufacturing technologies, and the manufacturing yields of the graphene membranes with suspended proof masses were greater than 90%, with more than 70% of the graphene membranes having more than 90% graphene area without visible defects. The graphene membranes with suspended proof masses were extremely robust and were able to withstand indentation forces from an atomic force microscope (AFM) tip of up to ~7000 nN. The measured resonance frequencies of the realized structures ranged from tens to hundreds of kHz, with quality factors ranging from 63 to 148. The proposed approach for the reliable and large-scale manufacture of graphene membranes with suspended proof masses will enable the development and study of innovative NEMS devices with new functionalities and improved performances.
△ Less
Submitted 16 March, 2020;
originally announced March 2020.
-
Independent Domination in Subcubic Graphs
Authors:
A. Akbari,
S. Akbari,
A. Doosthosseini,
Z. Hadizadeh,
Michael A. Henning,
A. Naraghi
Abstract:
A set $S$ of vertices in a graph $G$ is a dominating set if every vertex not in $S$ is adjacent to a vertex in $S$. If, in addition, $S$ is an independent set, then $S$ is an independent dominating set. The independent domination number $i(G)$ of $G$ is the minimum cardinality of an independent dominating set in $G$. In 2013 Goddard and Henning [Discrete Math 313 (2013), 839--854] conjectured that…
▽ More
A set $S$ of vertices in a graph $G$ is a dominating set if every vertex not in $S$ is adjacent to a vertex in $S$. If, in addition, $S$ is an independent set, then $S$ is an independent dominating set. The independent domination number $i(G)$ of $G$ is the minimum cardinality of an independent dominating set in $G$. In 2013 Goddard and Henning [Discrete Math 313 (2013), 839--854] conjectured that if $G$ is a connected cubic graph of order $n$, then $i(G) \le \frac{3}{8}n$, except if $G$ is the complete bipartite graph $K_{3,3}$ or the $5$-prism $C_5 \, \Box \, K_2$. Further they construct two infinite families of connected cubic graphs with independent domination three-eighths their order. They remark that perhaps it is even true that for $n > 10$ these two families are only families for which equality holds. In this paper, we provide a new family of connected cubic graphs $G$ of order $n$ such that $i(G) = \frac{3}{8}n$. We also show that if $G$ is a subcubic graph of order $n$ with no isolated vertex, then $i(G) \le \frac{1}{2}n$, and we characterize the graphs achieving equality in this bound.
△ Less
Submitted 9 January, 2020;
originally announced January 2020.
-
Some Results on Dominating Induced Matchings
Authors:
Saieed Akbari,
Hossein Baktash,
Amin Behjati,
Afshin Behmaram,
Mohammad Roghani
Abstract:
Let $G$ be a graph, a dominating induced matching (DIM) of $G$ is an induced matching that dominates every edge of $G$. In this paper we show that if a graph $G$ has a DIM, then $χ(G) \leqslant 3$. Also, it is shown that if $G$ is a connected graph whose all edges can be partitioned into DIM, then $G$ is either a regular graph or a biregular graph and indeed we characterize all graphs whose edge s…
▽ More
Let $G$ be a graph, a dominating induced matching (DIM) of $G$ is an induced matching that dominates every edge of $G$. In this paper we show that if a graph $G$ has a DIM, then $χ(G) \leqslant 3$. Also, it is shown that if $G$ is a connected graph whose all edges can be partitioned into DIM, then $G$ is either a regular graph or a biregular graph and indeed we characterize all graphs whose edge set can be partitioned into DIM. Also, we prove that if $G$ is an $r$-regular graph of order $n$ whose edges can be partitioned into DIM, then $n$ is divisible by $\binom{2r - 1}{r - 1}$ and $n = \binom{2r - 1}{r - 1}$ if and only if $G$ is the Kneser graph with parameters $r-1$, $2r-1$.
△ Less
Submitted 1 December, 2019;
originally announced December 2019.
-
A Note on Induced Path Decomposition of Graphs
Authors:
S. Akbari,
H. R. Maimani,
A. Seify
Abstract:
Let $G$ be a graph of order $n$. The path decomposition of $G$ is a set of disjoint paths, say $\mathcal{P}$, which cover all vertices of $G$. If all paths are induced paths in $G$, then we say $\mathcal{P}$ is an induced path decomposition of $G$. Moreover, if every path is of order at least 2, then we say $G$ has an IPD. In this paper, we prove that every connected $r$-regular graph which is not…
▽ More
Let $G$ be a graph of order $n$. The path decomposition of $G$ is a set of disjoint paths, say $\mathcal{P}$, which cover all vertices of $G$. If all paths are induced paths in $G$, then we say $\mathcal{P}$ is an induced path decomposition of $G$. Moreover, if every path is of order at least 2, then we say $G$ has an IPD. In this paper, we prove that every connected $r$-regular graph which is not complete graph of odd order admits an IPD. Also we show that every connected bipartite cubic graph of order $n$ admits an IPD of size at most $\frac{n}{3}$. We classify all connected claw-free graphs which admit an IPD.
△ Less
Submitted 1 December, 2019;
originally announced December 2019.
-
On the Maximum Order of Induced Paths and Induced Forests in Regular Graphs
Authors:
Saieed Akbari,
Alireza Amanihamedani,
Sepehr Mousavi,
Hesam Nikpey,
Soheil Sheybani
Abstract:
Let $G$ be a graph and $a(G)$, LIF$(G)$ denote the maximum orders of an induced forest and an induced linear forest of $G$, respectively. It is well-known that if $G$ is an $r$-regular graph of order $n$, then $a(G) \geq \frac{2}{r+1}n$. In this paper, we generalize this result by showing that LIF$(G) \geq \frac{2}{r+1}n$. It was proved that for every graph $G$,…
▽ More
Let $G$ be a graph and $a(G)$, LIF$(G)$ denote the maximum orders of an induced forest and an induced linear forest of $G$, respectively. It is well-known that if $G$ is an $r$-regular graph of order $n$, then $a(G) \geq \frac{2}{r+1}n$. In this paper, we generalize this result by showing that LIF$(G) \geq \frac{2}{r+1}n$. It was proved that for every graph $G$, $a(G) \geq \sum_{i=1}^{n}\frac{2}{d_i+1}$, where $d_1, \ldots, d_n$ is the degree sequence of $G$. Here, we conjecture that for every graph $G$ with $δ(G) \geq 2$, LIF$(G) \geq \sum_{i=1}^{n}\frac{2}{d_i+1}$.
△ Less
Submitted 6 November, 2019;
originally announced November 2019.
-
Trees with a large Laplacian eigenvalue multiplicity
Authors:
S. Akbari,
E. R. van Dam,
M. H. Fakharan
Abstract:
In this paper, we study the multiplicity of the Laplacian eigenvalues of trees. It is known that for trees, integer Laplacian eigenvalues larger than $1$ are simple and also the multiplicity of Laplacian eigenvalue $1$ has been well studied before. Here we consider the multiplicities of the other (non-integral) Laplacian eigenvalues. We give an upper bound and determine the trees of order $n$ that…
▽ More
In this paper, we study the multiplicity of the Laplacian eigenvalues of trees. It is known that for trees, integer Laplacian eigenvalues larger than $1$ are simple and also the multiplicity of Laplacian eigenvalue $1$ has been well studied before. Here we consider the multiplicities of the other (non-integral) Laplacian eigenvalues. We give an upper bound and determine the trees of order $n$ that have a multiplicity that is close to the upper bound $\frac{n-3}{2}$, and emphasize the particular role of the algebraic connectivity.
△ Less
Submitted 23 October, 2019; v1 submitted 26 July, 2019;
originally announced July 2019.
-
Star Coloring of the Cartesian Product of Cycles
Authors:
S Akbari,
M Chavooshi,
M Ghanbari,
S Taghian
Abstract:
A proper vertex coloring of a graph $G$ is called a star coloring if every two color classes induce a forest whose each component is a star, which means there is no bicolored $P_4$ in $G$. In this paper, we show that the Cartesian product of any two cycles, except $C_3 \square C_3$ and $C_3 \square C_5$, has a $5$-star coloring.
A proper vertex coloring of a graph $G$ is called a star coloring if every two color classes induce a forest whose each component is a star, which means there is no bicolored $P_4$ in $G$. In this paper, we show that the Cartesian product of any two cycles, except $C_3 \square C_3$ and $C_3 \square C_5$, has a $5$-star coloring.
△ Less
Submitted 15 June, 2019;
originally announced June 2019.
-
Well-indumatched Trees and Graphs of Bounded Girth
Authors:
S. Akbari,
T. Ekim,
A. H. Ghodrati,
S. Zare
Abstract:
A graph G is called well-indumatched if all of its maximal induced matchings have the same size. In this paper we characterize all well-indumatched trees. We provide a linear time algorithm to decide if a tree is well-indumatched or not. Then, we characterize minimal well-indumatched graphs of girth at least 9 and show subsequently that for an odd integer g greater than or equal to 9 and different…
▽ More
A graph G is called well-indumatched if all of its maximal induced matchings have the same size. In this paper we characterize all well-indumatched trees. We provide a linear time algorithm to decide if a tree is well-indumatched or not. Then, we characterize minimal well-indumatched graphs of girth at least 9 and show subsequently that for an odd integer g greater than or equal to 9 and different from 11, there is no well-indumatched graph of girth g. On the other hand, there are infinitely many well-indumatched unicyclic graphs of girth k, where k is in {3, 5, 7} or k is an even integer greater than 2. We also show that, although the recognition of well-indumatched graphs is known to be co-NP-complete in general, one can recognize in polynomial time well-indumatched graphs where the size of maximal induced matchings is fixed.
△ Less
Submitted 16 December, 2019; v1 submitted 7 March, 2019;
originally announced March 2019.
-
Proof of a Conjecture on the Seidel Energy of Graphs
Authors:
Saieed Akbari,
Mostafa Einollahzadeh,
Mohammad Mahdi Karkhaneei,
Mohammad Ali Nematollahi
Abstract:
Let $G$ be a graph with the vertex set $ \lbrace v_1,\ldots,v_n \rbrace$. The Seidel matrix of $G$ is an $n\times n$ matrix whose diagonal entries are zero, $ij$-th entry is $-1$ if $ v_{i} $ and $ v_{j} $ are adjacent and otherwise is $ 1 $. The Seidel energy of $G$ is defined to be the sum of absolute values of all eigenvalues of the Seidel matrix of $G$. Haemers conjectured that the Seidel ener…
▽ More
Let $G$ be a graph with the vertex set $ \lbrace v_1,\ldots,v_n \rbrace$. The Seidel matrix of $G$ is an $n\times n$ matrix whose diagonal entries are zero, $ij$-th entry is $-1$ if $ v_{i} $ and $ v_{j} $ are adjacent and otherwise is $ 1 $. The Seidel energy of $G$ is defined to be the sum of absolute values of all eigenvalues of the Seidel matrix of $G$. Haemers conjectured that the Seidel energy of any graph of order $n$ is at least $2n-2$ and, up to Seidel equivalence, the equality holds for $ K_{n} $. We establish the validity of Haemers' Conjecture in general.
△ Less
Submitted 20 January, 2019;
originally announced January 2019.
-
Social Status and Communication Behavior in an Evolving Social Network
Authors:
Sahand Akbari
Abstract:
The degree to which individuals can exert influence on propagation of information and opinion dynamics in online communities is highly dependent on their social status. Therefore, there is a high demand for identifying influential users in a community by predicting their social position in that community. Moreover, understanding how people with various social status behave, can shed light on the d…
▽ More
The degree to which individuals can exert influence on propagation of information and opinion dynamics in online communities is highly dependent on their social status. Therefore, there is a high demand for identifying influential users in a community by predicting their social position in that community. Moreover, understanding how people with various social status behave, can shed light on the dynamics of interaction in social networks. In this paper, I study an evolving online social network originated from an online community for university students and I tackle the problem of forecasting users' social status, represented as their PageRank, based on frequency of recurring temporal sequences of observed behavior, i.e. behavioral motifs. I show that individuals with different values of PageRank exhibit different behavior even in early weeks since the online community's inception and it is possible to forecast future PageRank values given frequency of behavioral motifs with high accuracy.
△ Less
Submitted 23 October, 2018;
originally announced October 2018.
-
Induced path factors of regular graphs
Authors:
Saieed Akbari,
Daniel Horsley,
Ian M. Wanless
Abstract:
An induced path factor of a graph $G$ is a set of induced paths in $G$ with the property that every vertex of $G$ is in exactly one of the paths. The induced path number $ρ(G)$ of $G$ is the minimum number of paths in an induced path factor of $G$. We show that if $G$ is a connected cubic graph on $n>6$ vertices, then $ρ(G)\le(n-1)/3$.
Fix an integer $k\ge3$. For each $n$, define…
▽ More
An induced path factor of a graph $G$ is a set of induced paths in $G$ with the property that every vertex of $G$ is in exactly one of the paths. The induced path number $ρ(G)$ of $G$ is the minimum number of paths in an induced path factor of $G$. We show that if $G$ is a connected cubic graph on $n>6$ vertices, then $ρ(G)\le(n-1)/3$.
Fix an integer $k\ge3$. For each $n$, define $\mathcal{M}_n$ to be the maximum value of $ρ(G)$ over all connected $k$-regular graphs $G$ on $n$ vertices. As $n\rightarrow\infty$ with $nk$ even, we show that $c_k=\lim(\mathcal{M}_n/n)$ exists. We prove that $5/18\le c_3\le1/3$ and $3/7\le c_4\le1/2$ and that $c_k=\frac12-O(k^{-1})$ for $k\rightarrow\infty$.
△ Less
Submitted 1 November, 2020; v1 submitted 12 September, 2018;
originally announced September 2018.
-
Decomposing Claw-free Subcubic Graphs and $4$-Chordal Subcubic Graphs
Authors:
Elham Aboomahigir,
Milad Ahanjideh,
Saieed Akbari
Abstract:
Hoffmann-Ostenhof's Conjecture states that the edge set of every connected cubic graph can be decomposed into a spanning tree, a matching and a $2$-regular subgraph. In this paper, we show that the conjecture holds for claw-free subcubic graphs and $4$-chordal subcubic graphs.
Hoffmann-Ostenhof's Conjecture states that the edge set of every connected cubic graph can be decomposed into a spanning tree, a matching and a $2$-regular subgraph. In this paper, we show that the conjecture holds for claw-free subcubic graphs and $4$-chordal subcubic graphs.
△ Less
Submitted 30 January, 2020; v1 submitted 28 June, 2018;
originally announced June 2018.
-
The Algebraic Connectivity of a Graph and its Complement
Authors:
B. Afshari,
S. Akbari,
M. J. Moghaddamzadeh,
B. Mohar
Abstract:
For a graph $G$, let $λ_2(G)$ denote its second smallest Laplacian eigenvalue. It was conjectured that $λ_2(G) + λ_2(\overline G) \ge 1$, where $\overline G$ is the complement of $G$. In this paper, it is shown that $\max\{λ_2(G), λ_2(\overline G)\} \ge 2/5$.
For a graph $G$, let $λ_2(G)$ denote its second smallest Laplacian eigenvalue. It was conjectured that $λ_2(G) + λ_2(\overline G) \ge 1$, where $\overline G$ is the complement of $G$. In this paper, it is shown that $\max\{λ_2(G), λ_2(\overline G)\} \ge 2/5$.
△ Less
Submitted 18 June, 2018;
originally announced June 2018.
-
Some Mixed Graphs Determined by Their Spectrum
Authors:
S. Akbari,
A. Ghafari,
M. Nahvi,
M. A. Nematollahi
Abstract:
A mixed graph is obtained from a graph by orienting some of its edges. The Hermitian adjacency matrix of a mixed graph with the vertex set $ \{v_{1}, \ldots , v_{n}\} $, is the matrix $ H=[h_{ij}]_{n \times n} $, where $ h_{ij}=-h_{ji}=i $ if there is a directed edge from $ v_{i} $ to $ v_{j} $, $ h_{ij}=1 $ if there exists an undirected edge between $v_i$ and $v_{j}$, and $h_{ij}=0$ otherwise. Th…
▽ More
A mixed graph is obtained from a graph by orienting some of its edges. The Hermitian adjacency matrix of a mixed graph with the vertex set $ \{v_{1}, \ldots , v_{n}\} $, is the matrix $ H=[h_{ij}]_{n \times n} $, where $ h_{ij}=-h_{ji}=i $ if there is a directed edge from $ v_{i} $ to $ v_{j} $, $ h_{ij}=1 $ if there exists an undirected edge between $v_i$ and $v_{j}$, and $h_{ij}=0$ otherwise. The Hermitian spectrum of a mixed graph is defined to be the spectrum of its Hermitian adjacency matrix. In this paper we study mixed graphs which are determined by their Hermitian spectrum (DHS). First, we show that each mixed cycle is switching equivalent to either a mixed cycle with no directed edges ($C_{n}$), a mixed cycle with exactly one directed edge ($C_{n}^{1}$), or a mixed cycle with exactly two consecutive directed edges with the same direction ($C_{n}^{2}$) and we determine the spectrum of these three types of cycles. Next, we characterize all DHS mixed paths and mixed cycles. We show that all mixed paths of even order, except $P_{8}$ and $P_{14}$, are DHS. It is also shown that mixed paths of odd order, except $P_{3}$, are not DHS. Also, all cospectral mates of $P_{8}$, $P_{14}$ and $P_{4k+1}$ and two families of cospectral mates of $P_{4k+3}$, where $k\geq1$, are introduced. Finally, we show that the mixed cycles $C_{2k}$ and $C_{2k}^{2}$, where $k\geq3$, are not DHS, but the mixed cycles $C_{4}$, $C_{4}^{2}$, $C_{2k+1}$, $C_{2k+1}^{2}$, $C_{2k+1}^{1}$ and $C_{2j}^{1}$ except $C_{7}^{1}$, $C_{9}^{1}$, $C_{12}^{1}$ and $C_{15}^{1}$, are DHS, where $k\geq1$ and $j\geq2$.
△ Less
Submitted 10 June, 2018;
originally announced June 2018.
-
Chromatic Number and Dichromatic Polynomial of Digraphs
Authors:
Saeed Akbari,
Amir Hossein Ghodrati,
Afrouz Jabalameli,
Morteza Saghafian
Abstract:
Let $G$ be a graph of order $n$. It is well-known that $α(G)\geq \sum_{i=1}^n \frac{1}{1+d_i}$, where $α(G)$ is the independence number of $G$ and $d_1,\ldots,d_n$ is the degree sequence of $G$. We extend this result to digraphs by showing that if $D$ is a digraph with $n$ vertices, then $ α(D)\geq \sum_{i=1}^n \left( \frac{1}{1+d_i^+} + \frac{1}{1+d_i^-}
- \frac{1}{1+d_i}\right)$, where $α(D)$…
▽ More
Let $G$ be a graph of order $n$. It is well-known that $α(G)\geq \sum_{i=1}^n \frac{1}{1+d_i}$, where $α(G)$ is the independence number of $G$ and $d_1,\ldots,d_n$ is the degree sequence of $G$. We extend this result to digraphs by showing that if $D$ is a digraph with $n$ vertices, then $ α(D)\geq \sum_{i=1}^n \left( \frac{1}{1+d_i^+} + \frac{1}{1+d_i^-}
- \frac{1}{1+d_i}\right)$, where $α(D)$ is the maximum size of an acyclic vertex set of $D$. Golowich proved that for any digraph $D$, $χ(D)\leq \lceil \frac{4k}{5} \rceil+2$, where $k=max(Δ^+(D),Δ^-(D))$. We give a short and simple proof for this result. Next, we investigate the chromatic number of tournaments and determine the unique tournament such that for every integer $k>1$, the number of proper $k$-colorings of that tournament is maximum among all strongly connected tournaments with the same number of vertices. Also, we find the chromatic polynomial of the strongly connected tournament with the minimum number of cycles.
△ Less
Submitted 16 November, 2017;
originally announced November 2017.
-
Signed graphs cospectral with the path
Authors:
Saieed Akbari,
Willem H. Haemers,
Hamid Reza Maimani,
Leila Parsaei Majd
Abstract:
A signed graph $Γ$ is said to be determined by its spectrum if every signed graph with the same spectrum as $Γ$ is switching isomorphic with $Γ$. Here it is proved that the path $P_n$, interpreted as a signed graph, is determined by its spectrum if and only if $n\equiv 0, 1$, or 2 (mod 4), unless $n\in\{8, 13, 14, 17, 29\}$, or $n=3$.
A signed graph $Γ$ is said to be determined by its spectrum if every signed graph with the same spectrum as $Γ$ is switching isomorphic with $Γ$. Here it is proved that the path $P_n$, interpreted as a signed graph, is determined by its spectrum if and only if $n\equiv 0, 1$, or 2 (mod 4), unless $n\in\{8, 13, 14, 17, 29\}$, or $n=3$.
△ Less
Submitted 10 May, 2018; v1 submitted 28 September, 2017;
originally announced September 2017.
-
Some Criteria for a Signed Graph to Have Full Rank
Authors:
S. Akbari,
A. Ghafari,
K. Kazemian,
M. Nahvi
Abstract:
A weighted graph $G^ω$ consists of a simple graph $G$ with a weight $ω$, which is a map**,$ω$: $E(G)\rightarrow\mathbb{Z}\backslash\{0\}$. A signed graph is a graph whose edges are labeled with $-1$ or $1$. In this paper, we characterize graphs which have a sign such that their signed adjacency matrix has full rank, and graphs which have a weight such that their weighted adjacency matrix does no…
▽ More
A weighted graph $G^ω$ consists of a simple graph $G$ with a weight $ω$, which is a map**,$ω$: $E(G)\rightarrow\mathbb{Z}\backslash\{0\}$. A signed graph is a graph whose edges are labeled with $-1$ or $1$. In this paper, we characterize graphs which have a sign such that their signed adjacency matrix has full rank, and graphs which have a weight such that their weighted adjacency matrix does not have full rank. We show that for any arbitrary simple graph $G$, there is a sign $σ$ so that $G^σ$ has full rank if and only if $G$ has a $\{1,2\}$-factor. We also show that for a graph $G$, there is a weight $ω$ so that $G^ω$ does not have full rank if and only if $G$ has at least two $\{1,2\}$-factors.
△ Less
Submitted 23 August, 2017;
originally announced August 2017.
-
Maximum nullity and zero forcing number on cubic graphs
Authors:
Saieed Akbari,
Ebrahim Vatandoost,
Yasser Golkhandy Pour
Abstract:
Let $G$ be a graph. The maximum nullity of $G$, denoted by $M(G)$, is defined to be the largest possible nullity over all real symmetric matrices $A$ whose $a_{ij}\neq 0$ for $i\neq j$, whenever two vertices $u_i$ and $u_j$ of $G$ are adjacent. In this paper, we characterize all cubic graphs with zero forcing number $3$. As a corollary, it is shown that if the zero forcing number is $3$, then…
▽ More
Let $G$ be a graph. The maximum nullity of $G$, denoted by $M(G)$, is defined to be the largest possible nullity over all real symmetric matrices $A$ whose $a_{ij}\neq 0$ for $i\neq j$, whenever two vertices $u_i$ and $u_j$ of $G$ are adjacent. In this paper, we characterize all cubic graphs with zero forcing number $3$. As a corollary, it is shown that if the zero forcing number is $3$, then $M(G)=3$. In addition, we introduce a family of cubic graphs containing graphs $G$ with $M(G)=Z(G)=4$.
Also, we provide an algorithm which make a relation between maximum nullity of $G$ and the number of leaves in a spanning tree of $G$.
△ Less
Submitted 27 May, 2017;
originally announced May 2017.
-
Graphs with Integer Matching Polynomial Roots
Authors:
S. Akbari,
P. Csikvari,
A. Ghafari,
S. Khalashi Ghezelahmad,
M. Nahvi
Abstract:
In this paper, we study graphs whose matching polynomial have only integer zeros. A graph is matching integral if the zeros of its matching polynomial are all integers. We characterize all matching integral traceable graphs.. We show that apart from K7 n (E(C3) [ E(C4)) there is no connected k-regular matching integral graph if k ? 2. It is also shown that if G is a graph with a perfect matching,…
▽ More
In this paper, we study graphs whose matching polynomial have only integer zeros. A graph is matching integral if the zeros of its matching polynomial are all integers. We characterize all matching integral traceable graphs.. We show that apart from K7 n (E(C3) [ E(C4)) there is no connected k-regular matching integral graph if k ? 2. It is also shown that if G is a graph with a perfect matching, then its matching polynomial has a zero in the interval (0, 1]. Finally, we describe all claw-free matching integral graphs.
△ Less
Submitted 5 February, 2017; v1 submitted 2 August, 2016;
originally announced August 2016.
-
Hoffmann-Ostenhof's conjecture for traceable cubic graphs
Authors:
F. Abdolhosseini,
S. Akbari,
H. Hashemi,
M. S. Moradian
Abstract:
It was conjectured by Hoffmann-Ostenhof that the edge set of every connected cubic graph can be decomposed into a spanning tree, a matching and a family of cycles. In this paper, we show that this conjecture holds for traceable cubic graphs.
It was conjectured by Hoffmann-Ostenhof that the edge set of every connected cubic graph can be decomposed into a spanning tree, a matching and a family of cycles. In this paper, we show that this conjecture holds for traceable cubic graphs.
△ Less
Submitted 16 July, 2016;
originally announced July 2016.
-
Equimatchable Claw-Free Graphs
Authors:
Saieed Akbari,
Hadi Alizadeh,
Tınaz Ekim,
Didem Gözüpek,
Mordechai Shalom
Abstract:
A graph is equimatchable if all of its maximal matchings have the same size. A graph is claw-free if it does not have a claw as an induced subgraph. In this paper, we provide, to the best of our knowledge, the first characterization of claw-free equimatchable graphs by identifying the equimatchable claw-free graph families. This characterization implies an efficient recognition algorithm.
A graph is equimatchable if all of its maximal matchings have the same size. A graph is claw-free if it does not have a claw as an induced subgraph. In this paper, we provide, to the best of our knowledge, the first characterization of claw-free equimatchable graphs by identifying the equimatchable claw-free graph families. This characterization implies an efficient recognition algorithm.
△ Less
Submitted 24 July, 2018; v1 submitted 2 July, 2016;
originally announced July 2016.
