-
Multi-Period Max Flow Network Interdiction with Restructuring for Disrupting Domestic Sex Trafficking Networks
Authors:
Daniel Kosmas,
Thomas C Sharkey,
John E Mitchell,
Kayse Lee Maass,
Lauren Martin
Abstract:
We consider a new class of multi-period network interdiction problems, where interdiction and restructuring decisions are decided upon before the network is operated and implemented throughout the time horizon. We discuss how we apply this new problem to disrupting domestic sex trafficking networks, and introduce a variant where a second cooperating attacker has the ability to interdict victims an…
▽ More
We consider a new class of multi-period network interdiction problems, where interdiction and restructuring decisions are decided upon before the network is operated and implemented throughout the time horizon. We discuss how we apply this new problem to disrupting domestic sex trafficking networks, and introduce a variant where a second cooperating attacker has the ability to interdict victims and prevent the recruitment of prospective victims. This problem is modeled as a bilevel mixed integer linear program (BMILP), and is solved using column-and-constraint generation with partial information. We also simplify the BMILP when all interdictions are implemented before the network is operated. Modeling-based augmentations are proposed to significantly improve the solution time in a majority of instances tested. We apply our method to synthetic domestic sex trafficking networks, and discuss policy implications from our model. In particular, we show how preventing the recruitment of prospective victims may be as essential to disrupting sex trafficking as interdicting existing participants.
△ Less
Submitted 2 December, 2022; v1 submitted 9 September, 2022;
originally announced September 2022.
-
Provable Low Rank Plus Sparse Matrix Separation Via Nonconvex Regularizers
Authors:
April Sagan,
John E. Mitchell
Abstract:
This paper considers a large class of problems where we seek to recover a low rank matrix and/or sparse vector from some set of measurements. While methods based on convex relaxations suffer from a (possibly large) estimator bias, and other nonconvex methods require the rank or sparsity to be known a priori, we use nonconvex regularizers to minimize the rank and $l_0$ norm without the estimator bi…
▽ More
This paper considers a large class of problems where we seek to recover a low rank matrix and/or sparse vector from some set of measurements. While methods based on convex relaxations suffer from a (possibly large) estimator bias, and other nonconvex methods require the rank or sparsity to be known a priori, we use nonconvex regularizers to minimize the rank and $l_0$ norm without the estimator bias from the convex relaxation. We present a novel analysis of the alternating proximal gradient descent algorithm applied to such problems, and bound the error between the iterates and the ground truth sparse and low rank matrices. The algorithm and error bound can be applied to sparse optimization, matrix completion, and robust principal component analysis as special cases of our results.
△ Less
Submitted 26 September, 2021;
originally announced September 2021.
-
Precompensation of 3D field distortions in remote focus two-photon microscopy
Authors:
Antoine M. Valera,
Fiona C. Neufeldt,
Paul A. Kirkby,
John E. Mitchell,
R. Angus Silver
Abstract:
Remote focusing is widely used in 3D two-photon microscopy and 3D photostimulation because it enables fast axial scanning without moving the objective lens or specimen. However, due to the design constraints of microscope optics, remote focus units are often located in non-telecentric positions in the optical path, leading to significant depth dependent 3D field distortions in the imaging volume.…
▽ More
Remote focusing is widely used in 3D two-photon microscopy and 3D photostimulation because it enables fast axial scanning without moving the objective lens or specimen. However, due to the design constraints of microscope optics, remote focus units are often located in non-telecentric positions in the optical path, leading to significant depth dependent 3D field distortions in the imaging volume. To address this limitation, we characterized 3D field distortions arising from non-telecentric remote focusing and present a method for distortion precompensation. We demonstrate its applicability for a 3D two-photon microscope that uses an acousto-optic lens (AOL) for remote focusing and scanning. We show that the distortion precompensation method improves the pointing precision of the AOL microscope to < 0.5 micrometer throughout the 400 x 400 x 400 micrometer imaging volume.
△ Less
Submitted 18 March, 2021;
originally announced March 2021.
-
Optimizing Edge Sets in Networks to Produce Ground Truth Communities Based on Modularity
Authors:
Daniel Kosmas,
John E. Mitchell,
Thomas C. Sharkey,
Boleslaw K. Szymanski
Abstract:
We consider two new problems regarding the impact of edge addition or removal on the modularity of partitions (or community structures) in a network. The first problem seeks to add edges to enforce that a desired partition is the partition that maximizes modularity. The second problem seeks to find the sparsest representation of a network that has the same partition with maximum modularity as the…
▽ More
We consider two new problems regarding the impact of edge addition or removal on the modularity of partitions (or community structures) in a network. The first problem seeks to add edges to enforce that a desired partition is the partition that maximizes modularity. The second problem seeks to find the sparsest representation of a network that has the same partition with maximum modularity as the original network. We present integer programming formulations, a row generation algorithm, and heuristic algorithms to solve these problems. Further, we demonstrate a counter-intuitive behavior of modularity that makes the development of heuristics for general networks difficult. We then present results on a selection of social and illicit networks from the literature.
△ Less
Submitted 3 March, 2022; v1 submitted 16 March, 2021;
originally announced March 2021.
-
Interdicting Restructuring Networks with Applications in Illicit Trafficking
Authors:
Daniel Kosmas,
Thomas C. Sharkey,
John E. Mitchell,
Kayse Lee Maass,
Lauren Martin
Abstract:
We consider a new class of max flow network interdiction problems, where the defender is able to introduce new arcs to the network after the attacker has made their interdiction decisions. We prove properties of when this restructuring will not increase the value of the minimum cut, which has important practical interpretations for problems of disrupting drug trafficking networks. In particular, i…
▽ More
We consider a new class of max flow network interdiction problems, where the defender is able to introduce new arcs to the network after the attacker has made their interdiction decisions. We prove properties of when this restructuring will not increase the value of the minimum cut, which has important practical interpretations for problems of disrupting drug trafficking networks. In particular, it demonstrates that disrupting lower levels of these networks will not impact their operations when replacing the disrupted participants is easy. For the bilevel mixed integer linear programming formulation of this problem, we devise a column-and-constraint generation (C&CG) algorithm to solve it. Our approach uses partial information on the feasibility of restructuring plans and is shown to be orders of magnitude faster than previous C&CG methods. We demonstrate that applying decisions from standard max flow network interdiction problems can result in significantly higher flows than interdictions that account for the restructuring.
△ Less
Submitted 2 December, 2022; v1 submitted 13 November, 2020;
originally announced November 2020.
-
Training Deep Neural Networks with Constrained Learning Parameters
Authors:
Prasanna Date,
Christopher D. Carothers,
John E. Mitchell,
James A. Hendler,
Malik Magdon-Ismail
Abstract:
Today's deep learning models are primarily trained on CPUs and GPUs. Although these models tend to have low error, they consume high power and utilize large amount of memory owing to double precision floating point learning parameters. Beyond the Moore's law, a significant portion of deep learning tasks would run on edge computing systems, which will form an indispensable part of the entire comput…
▽ More
Today's deep learning models are primarily trained on CPUs and GPUs. Although these models tend to have low error, they consume high power and utilize large amount of memory owing to double precision floating point learning parameters. Beyond the Moore's law, a significant portion of deep learning tasks would run on edge computing systems, which will form an indispensable part of the entire computation fabric. Subsequently, training deep learning models for such systems will have to be tailored and adopted to generate models that have the following desirable characteristics: low error, low memory, and low power. We believe that deep neural networks (DNNs), where learning parameters are constrained to have a set of finite discrete values, running on neuromorphic computing systems would be instrumental for intelligent edge computing systems having these desirable characteristics. To this extent, we propose the Combinatorial Neural Network Training Algorithm (CoNNTrA), that leverages a coordinate gradient descent-based approach for training deep learning models with finite discrete learning parameters. Next, we elaborate on the theoretical underpinnings and evaluate the computational complexity of CoNNTrA. As a proof of concept, we use CoNNTrA to train deep learning models with ternary learning parameters on the MNIST, Iris and ImageNet data sets and compare their performance to the same models trained using Backpropagation. We use following performance metrics for the comparison: (i) Training error; (ii) Validation error; (iii) Memory usage; and (iv) Training time. Our results indicate that CoNNTrA models use 32x less memory and have errors at par with the Backpropagation models.
△ Less
Submitted 1 September, 2020;
originally announced September 2020.
-
Low-Rank Factorization for Rank Minimization with Nonconvex Regularizers
Authors:
April Sagan,
John E. Mitchell
Abstract:
Rank minimization is of interest in machine learning applications such as recommender systems and robust principal component analysis. Minimizing the convex relaxation to the rank minimization problem, the nuclear norm, is an effective technique to solve the problem with strong performance guarantees. However, nonconvex relaxations have less estimation bias than the nuclear norm and can more accur…
▽ More
Rank minimization is of interest in machine learning applications such as recommender systems and robust principal component analysis. Minimizing the convex relaxation to the rank minimization problem, the nuclear norm, is an effective technique to solve the problem with strong performance guarantees. However, nonconvex relaxations have less estimation bias than the nuclear norm and can more accurately reduce the effect of noise on the measurements.
We develop efficient algorithms based on iteratively reweighted nuclear norm schemes, while also utilizing the low rank factorization for semidefinite programs put forth by Burer and Monteiro. We prove convergence and computationally show the advantages over convex relaxations and alternating minimization methods. Additionally, the computational complexity of each iteration of our algorithm is on par with other state of the art algorithms, allowing us to quickly find solutions to the rank minimization problem for large matrices.
△ Less
Submitted 28 March, 2021; v1 submitted 13 June, 2020;
originally announced June 2020.
-
Solving Linear Programs with Complementarity Constraints using Branch-and-Cut
Authors:
Bin Yu,
John E. Mitchell,
Jong-Shi Pang
Abstract:
A linear program with linear complementarity constraints (LPCC) requires the minimization of a linear objective over a set of linear constraints together with additional linear complementarity constraints. This class has emerged as a modeling paradigm for a broad collection of problems, including bilevel programs, Stackelberg games, inverse quadratic programs, and problems involving equilibrium co…
▽ More
A linear program with linear complementarity constraints (LPCC) requires the minimization of a linear objective over a set of linear constraints together with additional linear complementarity constraints. This class has emerged as a modeling paradigm for a broad collection of problems, including bilevel programs, Stackelberg games, inverse quadratic programs, and problems involving equilibrium constraints. The presence of the complementarity constraints results in a nonconvex optimization problem. We develop a branch-and-cut algorithm to find a global optimum for this class of optimization problems, where we branch directly on complementarities. We develop branching rules and feasibility recovery procedures and demonstrate their computational effectiveness in a comparison with CPLEX. The implementation builds on CPLEX through the use of callback routines. The computational results show that our approach is a strong alternative to constructing an integer programming formulation using big-$M$ terms to represent bounds for variables, with testing conducted on general LPCCs as well as on instances generated from bilevel programs with convex quadratic lower level problems.
△ Less
Submitted 8 February, 2018;
originally announced February 2018.
-
A Penalty Method for Rank Minimization Problems in Symmetric Matrices
Authors:
Xin Shen,
John E. Mitchell
Abstract:
The problem of minimizing the rank of a symmetric positive semidefinite matrix subject to constraints can be cast equivalently as a semidefinite program with complementarity constraints (SDCMPCC). The formulation requires two positive semidefinite matrices to be complementary. This is a continuous and nonconvex reformulation of the rank minimization problem. We investigate calmness of locally opti…
▽ More
The problem of minimizing the rank of a symmetric positive semidefinite matrix subject to constraints can be cast equivalently as a semidefinite program with complementarity constraints (SDCMPCC). The formulation requires two positive semidefinite matrices to be complementary. This is a continuous and nonconvex reformulation of the rank minimization problem. We investigate calmness of locally optimal solutions to the SDCMPCC formulation and hence show that any locally optimal solution is a KKT point.
We develop a penalty formulation of the problem. We present calmness results for locally optimal solutions to the penalty formulation. We also develop a proximal alternating linearized minimization (PALM) scheme for the penalty formulation, and investigate the incorporation of a momentum term into the algorithm. Computational results are presented.
△ Less
Submitted 1 February, 2018; v1 submitted 11 January, 2017;
originally announced January 2017.
-
Optimization of a Non-arsenic Iron-based Superconductor for Wire Fabrication
Authors:
Jonathan E. Mitchell,
Daniel A. Hillesheim,
Craig A. Bridges,
M. Parans Paranthaman,
Kris Gofryk,
Mike Rindfleisch,
Mike Tomsic,
Athena S. Sefat
Abstract:
We report on the optimization of synthesis of iron-selenide (non-arsenic) superconducting powders that are based on '122' composition, with optimal Tc = 38 K and Jc = 10^5 A/cm2 (4 K). We also report on the wire proof-of concept for these materials, by producing ~ 40 ft of wire that produce Ic. The 122 selenides are more difficult to synthesize and have more complex crystal structures compared to…
▽ More
We report on the optimization of synthesis of iron-selenide (non-arsenic) superconducting powders that are based on '122' composition, with optimal Tc = 38 K and Jc = 10^5 A/cm2 (4 K). We also report on the wire proof-of concept for these materials, by producing ~ 40 ft of wire that produce Ic. The 122 selenides are more difficult to synthesize and have more complex crystal structures compared to '11' selenides (FeSe and FeSe1-xTex), but they do offer higher Tc and might provoke a natural extension for 11 work.
△ Less
Submitted 9 December, 2014;
originally announced December 2014.
-
Orbital occupancy and charge do** in iron-based superconductors
Authors:
Claudia Cantoni,
Jonathan E. Mitchell,
Andrew F. May,
Michael A. McGuire,
Juan-Carlos Idrobo,
Tom Berlijn,
Elbio Dagotto,
Matthew F. Chisholm,
Wu Zhou,
Stephen J. Pennycook,
Athena S. Sefat,
Brian C. Sales
Abstract:
Iron-based superconductors (FBS) comprise several families of compounds having the same atomic building blocks for superconductivity, but large discrepancies among their physical properties. A longstanding goal in the field has been to decipher the key underlying factors controlling TC and the various do** mechanisms. In FBS materials this is complicated immensely by the different crystal and ma…
▽ More
Iron-based superconductors (FBS) comprise several families of compounds having the same atomic building blocks for superconductivity, but large discrepancies among their physical properties. A longstanding goal in the field has been to decipher the key underlying factors controlling TC and the various do** mechanisms. In FBS materials this is complicated immensely by the different crystal and magnetic structures exhibited by the different families. In this paper, using aberration-corrected scanning transmission electron microscopy (STEM) coupled with electron energy loss spectroscopy (EELS), we observe a universal behavior in the hole concentration and magnetic moment across different families. All the parent materials have the same total number of electrons in the Fe 3d bands; however, the local Fe magnetic moment varies due to different orbital occupancy. Although the common understanding has been that both long-range and local magnetic moments decrease with do**, we find that, near the onset of superconductivity, the local magnetic moment increases and shows a dome-like maximum near optimal do**, where no ordered magnetic moment is present. In addition, we address a longstanding debate concerning how Co substitutions induces superconductivity in the 122 arsenide family, showing that the 3d band filling increases a function of do**. These new microscopic insights into the properties of FBS demonstrate the importance of spin fluctuations for the superconducting state, reveal changes in orbital occupancy among different families of FBS, and confirm charge do** as one of the mechanisms responsible for superconductivity in 122 arsenides.
△ Less
Submitted 3 October, 2014;
originally announced October 2014.
-
Local inhomogeneity and filamentary superconductivity in Pr-doped CaFe$_{2}$As$_{2}$
Authors:
Krzysztof Gofryk,
Minghu Pan,
Claudia Cantoni,
Bayrammurad Saparov,
Jonathan E. Mitchell,
Athena S. Sefat
Abstract:
We use multi-scale techniques to determine the extent of local inhomogeneity and superconductivity in Ca$_{0.86}$Pr$_{0.14}$Fe$_{2}$As$_{2}$ single crystal. The inhomogeneity is manifested as a spatial variation of praseodymium concentration, local density of states, and superconducting order parameter. We show that the high-$T_{c}$ superconductivity emerges from clover-like defects associated wit…
▽ More
We use multi-scale techniques to determine the extent of local inhomogeneity and superconductivity in Ca$_{0.86}$Pr$_{0.14}$Fe$_{2}$As$_{2}$ single crystal. The inhomogeneity is manifested as a spatial variation of praseodymium concentration, local density of states, and superconducting order parameter. We show that the high-$T_{c}$ superconductivity emerges from clover-like defects associated with Pr dopants. The highest $T_{c}$ is observed in both the tetragonal and collapsed tetragonal phases, and its filamentary nature is a consequence of non-uniform Pr distribution that develops localized, isolated superconducting regions within the crystals.
△ Less
Submitted 7 January, 2014;
originally announced January 2014.
-
The influence of spin fluctuations on the thermal conductivity in superconducting Ba(Fe_{1-x}Co_x)_2As_2
Authors:
Andrew F. May,
Michael A. McGuire,
Jonathan E. Mitchell,
Athena S. Sefat,
Brian C. Sales
Abstract:
The thermal conductivity of electron-doped Ba(Fe$_{1-x}$Co$_x$)$_2$As$_2$ single crystals is investigated below 200K, with an emphasis on the behavior near the magnetic and superconducting (T_c) transition temperatures. An enhancement of the in-plane thermal conductivity $κ_{ab}$ is observed below T_c for all samples, with the greatest enhancement observed near optimal do**. The observed trends…
▽ More
The thermal conductivity of electron-doped Ba(Fe$_{1-x}$Co$_x$)$_2$As$_2$ single crystals is investigated below 200K, with an emphasis on the behavior near the magnetic and superconducting (T_c) transition temperatures. An enhancement of the in-plane thermal conductivity $κ_{ab}$ is observed below T_c for all samples, with the greatest enhancement observed near optimal do**. The observed trends are consistent with the scattering of heat carriers by low-energy magnetic excitations. Upon entering the superconducting state, the formation of a spin-gap leads to reduced scattering and an enhancement in $κ(T)$. Similarly, an enhancement of $κ$ is observed for polycrystalline BaFe2As2 below the magnetic transition, and qualitative differences in $κ(T)$ between single crystalline and polycrystalline BaFe2As2 are utilized to discuss anisotropic scattering. This study highlights how measuring $κ$ near $T_c$ in novel superconductors can be useful as a means to probe the potential role of spin fluctuations.
△ Less
Submitted 8 August, 2013;
originally announced August 2013.
-
Temperature-composition Phase Diagrams for Ba1-xSrxFe2As2 and Ba0.5Sr0.5(Fe1-yCoy)2As2
Authors:
Jonathan E. Mitchell,
Bayrammurad Saparov,
Wenzhi Lin,
Stuart Calder,
Qing Li,
Sergei V. Kalinin,
Minghu Pan,
Andrew D. Christianson,
Athena S. Sefat
Abstract:
Single crystals of mixed alkaline earth metal iron arsenide materials of Ba1-xSrxFe2As2 and Ba0.5Sr0.5(Fe1-yCoy)2As2 are synthesized via the self-flux method. Ba1-xSrxFe2As2 display spin-density wave features (TN) at temperatures intermediate to the parent materials, x = 0 and 1, with TN(x) following an approximately linear trend. Cobalt do** of the 1 to 1 Ba:Sr mixture, Ba0.5Sr0.5(Fe1-yCoy)2As2…
▽ More
Single crystals of mixed alkaline earth metal iron arsenide materials of Ba1-xSrxFe2As2 and Ba0.5Sr0.5(Fe1-yCoy)2As2 are synthesized via the self-flux method. Ba1-xSrxFe2As2 display spin-density wave features (TN) at temperatures intermediate to the parent materials, x = 0 and 1, with TN(x) following an approximately linear trend. Cobalt do** of the 1 to 1 Ba:Sr mixture, Ba0.5Sr0.5(Fe1-yCoy)2As2, results in a superconducting dome with maximum transition temperature of TC = 19 K at y = 0.092, close to the maximum transition temperatures observed in unmixed A(Fe1-yCoy)2As2; however, an annealed crystal with y = 0.141 showed a TC increase from 11 to 16 K with a decrease in Sommerfeld coefficient from 2.58(2) to 0.63(2) mJ/(K2 mol atom). For the underdoped y = 0.053, neutron diffraction results give evidence that TN and structural transition (To) are linked at 78 K, with anomalies observed in magnetization, resistivity and heat capacity data, while a superconducting transition at TC ~ 6 K is seen in resistivity and heat capacity data. Scanning tunneling microscopy measurements for y = 0.073 give Dynes broadening factor of 1.15 and a superconducting gap of 2.37 meV with evidence of surface inhomogeneity.
△ Less
Submitted 24 October, 2012;
originally announced October 2012.
-
Properties of Binary Transition-Metal Arsenides (TAs)
Authors:
B. Saparov,
J. E. Mitchell,
A. S. Sefat
Abstract:
We present thermodynamic and transport properties of transition-metal (T) arsenides, TAs with T = Sc to Ni (3d), Zr, Nb, Ru (4d), Hf and Ta (5d). Characterization of these binaries is made with powder X-ray diffraction, temperature and field-dependent magnetization and resistivity, temperature-dependent heat capacity, Seebeck coefficient, and thermal conductivity. All binaries show metallic behavi…
▽ More
We present thermodynamic and transport properties of transition-metal (T) arsenides, TAs with T = Sc to Ni (3d), Zr, Nb, Ru (4d), Hf and Ta (5d). Characterization of these binaries is made with powder X-ray diffraction, temperature and field-dependent magnetization and resistivity, temperature-dependent heat capacity, Seebeck coefficient, and thermal conductivity. All binaries show metallic behavior except TaAs and RuAs. TaAs, NbAs, ScAs and ZrAs are diamagnetic, while CoAs, VAs, TiAs, NiAs and RuAs show approximately Pauli paramagnetic behavior. FeAs and CrAs undergo antiferromagnetic order below TN = 71 K and TN \approx 260 K, respectively. MnAs is a ferromagnet below TC = 317 K and undergoes hexagonal-orthorhombic-hexagonal transitions at TS = 317 K and 384 K, respectively. For TAs, Seebeck coefficients vary between + 40 uV/K and - 40 uV/K in the 2 K to 300 K range, whereas thermal conductivity values stay below 18 W/(m K). The Sommerfeld-coefficient γ are less than 10 mJ/(K2mol). At room temperature with application of 8 Tesla magnetic field, large positive magnetoresistance is found for TaAs (~25%), MnAs (~90%) and for NbAs (~75%).
△ Less
Submitted 17 May, 2012;
originally announced May 2012.
