Showing 1–14 of 14 results for author: Aspman, J

Robust Quantum Gate Complexity: Foundations

Authors: Johannes Aspman, Vyacheslav Kungurtsev, Jakub Marecek

Abstract: Optimal control of closed quantum systems is a well studied geometrically elegant set of computational theory and techniques that have proven pivotal in the implementation and understanding of quantum computers. The design of a circuit itself corresponds to an optimal control problem of choosing the appropriate set of gates (which appear as control operands) in order to steer a qubit from an initi… ▽ More Optimal control of closed quantum systems is a well studied geometrically elegant set of computational theory and techniques that have proven pivotal in the implementation and understanding of quantum computers. The design of a circuit itself corresponds to an optimal control problem of choosing the appropriate set of gates (which appear as control operands) in order to steer a qubit from an initial, easily prepared state, to one that is informative to the user in some sense, for e.g., an oracle whose evaluation is part of the circuit. However, contemporary devices are known to be noisy, and it is not certain that a circuit will behave as intended. Yet, although the computational tools exist in broader optimal control theory, robustness of adequate operation of a quantum control system with respect to uncertainty and errors has not yet been broadly studied in the literature. In this paper, we propose a new approach inspired by the closed quantum optimal control and its connection to geometric interpretations. To this end, we present the appropriate problem definitions of robustness in the context of quantum control, focusing on its broader implications for gate complexity. △ Less

Submitted 26 April, 2024; v1 submitted 24 April, 2024; originally announced April 2024.

Topological twists of massive SQCD, Part II

Authors: Johannes Aspman, Elias Furrer, Jan Manschot

Abstract: This is the second and final part of ``Topological twists of massive SQCD''. Part I is available at arXiv:2206.08943. In this second part, we evaluate the contribution of the Coulomb branch to topological path integrals for $\mathcal{N}=2$ supersymmetric QCD with $N_f\leq 3$ massive hypermultiplets on compact four-manifolds. Our analysis includes the decoupling of hypermultiplets, the massless lim… ▽ More This is the second and final part of ``Topological twists of massive SQCD''. Part I is available at arXiv:2206.08943. In this second part, we evaluate the contribution of the Coulomb branch to topological path integrals for $\mathcal{N}=2$ supersymmetric QCD with $N_f\leq 3$ massive hypermultiplets on compact four-manifolds. Our analysis includes the decoupling of hypermultiplets, the massless limit and the merging of mutually non-local singularities at the Argyres-Douglas points. We give explicit mass expansions for the four-manifolds $\mathbb{P}^2$ and $K3$. For $\mathbb{P}^2$, we find that the correlation functions are polynomial as function of the masses, while infinite series and (potential) singularities occur for $K3$. The mass dependence corresponds mathematically to the integration of the equivariant Chern class of the matter bundle over the moduli space of $Q$-fixed equations. We demonstrate that the physical partition functions agree with mathematical results on Segre numbers of instanton moduli spaces. △ Less

Submitted 18 December, 2023; originally announced December 2023.

Comments: 80 pages + appendices, 13 figures. Second part of a series of two papers, first part is available at arXiv:2206.08943

Piecewise Polynomial Regression of Tame Functions via Integer Programming

Authors: Gilles Bareilles, Johannes Aspman, Jiri Nemecek, Jakub Marecek

Abstract: Tame functions are a class of nonsmooth, nonconvex functions, which feature in a wide range of applications: functions encountered in the training of deep neural networks with all common activations, value functions of mixed-integer programs, or wave functions of small molecules. We consider approximating tame functions with piecewise polynomial functions. We bound the quality of approximation of… ▽ More Tame functions are a class of nonsmooth, nonconvex functions, which feature in a wide range of applications: functions encountered in the training of deep neural networks with all common activations, value functions of mixed-integer programs, or wave functions of small molecules. We consider approximating tame functions with piecewise polynomial functions. We bound the quality of approximation of a tame function by a piecewise polynomial function with a given number of segments on any full-dimensional cube. We also present the first mixed-integer programming formulation of piecewise polynomial regression. Together, these can be used to estimate tame functions. We demonstrate promising computational results. △ Less

Submitted 4 June, 2024; v1 submitted 22 November, 2023; originally announced November 2023.

Taming Binarized Neural Networks and Mixed-Integer Programs

Authors: Johannes Aspman, Georgios Korpas, Jakub Marecek

Abstract: There has been a great deal of recent interest in binarized neural networks, especially because of their explainability. At the same time, automatic differentiation algorithms such as backpropagation fail for binarized neural networks, which limits their applicability. By reformulating the problem of training binarized neural networks as a subadditive dual of a mixed-integer program, we show that… ▽ More There has been a great deal of recent interest in binarized neural networks, especially because of their explainability. At the same time, automatic differentiation algorithms such as backpropagation fail for binarized neural networks, which limits their applicability. By reformulating the problem of training binarized neural networks as a subadditive dual of a mixed-integer program, we show that binarized neural networks admit a tame representation. This, in turn, makes it possible to use the framework of Bolte et al. for implicit differentiation, which offers the possibility for practical implementation of backpropagation in the context of binarized neural networks. This approach could also be used for a broader class of mixed-integer programs, beyond the training of binarized neural networks, as encountered in symbolic approaches to AI and beyond. △ Less

Submitted 20 December, 2023; v1 submitted 5 October, 2023; originally announced October 2023.

Comments: 9 pages, 4 figures

Journal ref: Proceedings of the AAAI Conference on Artificial Intelligence, 2024

Approaching Collateral Optimization for NISQ and Quantum-Inspired Computing

Authors: Megan Giron, Georgios Korpas, Waqas Parvaiz, Prashant Malik, Johannes Aspman

Abstract: Collateral optimization refers to the systematic allocation of financial assets to satisfy obligations or secure transactions, while simultaneously minimizing costs and optimizing the usage of available resources. {This involves assessing number of characteristics, such as cost of funding and quality of the underlying assets to ascertain the optimal collateral quantity to be posted to cover exposu… ▽ More Collateral optimization refers to the systematic allocation of financial assets to satisfy obligations or secure transactions, while simultaneously minimizing costs and optimizing the usage of available resources. {This involves assessing number of characteristics, such as cost of funding and quality of the underlying assets to ascertain the optimal collateral quantity to be posted to cover exposure arising from a given transaction or a set of transactions. One of the common objectives is to minimise the cost of collateral required to mitigate the risk associated with a particular transaction or a portfolio of transactions while ensuring sufficient protection for the involved parties}. Often, this results in a large-scale combinatorial optimization problem. In this study, we initially present a Mixed Integer Linear Programming (MILP) formulation for the collateral optimization problem, followed by a Quadratic Unconstrained Binary optimization (QUBO) formulation in order to pave the way towards approaching the problem in a hybrid-quantum and NISQ-ready way. We conduct local computational small-scale tests using various Software Development Kits (SDKs) and discuss the behavior of our formulations as well as the potential for performance enhancements. We further survey the recent literature that proposes alternative ways to attack combinatorial optimization problems suitable for collateral optimization. △ Less

Submitted 19 December, 2023; v1 submitted 25 May, 2023; originally announced May 2023.

Comments: 14 pages, 4 figures, Published at IEEE Transactions on Quantum Engineering

Journal ref: in IEEE Transactions on Quantum Engineering, vol. 4, pp. 1-18, 2023, Art no. 3101818

Hybrid Methods in Polynomial Optimisation

Authors: Johannes Aspman, Gilles Bareilles, Vyacheslav Kungurtsev, Jakub Marecek, Martin Takáč

Abstract: The Moment/Sum-of-squares hierarchy provides a way to compute the global minimizers of polynomial optimization problems (POP), at the cost of solving a sequence of increasingly large semidefinite programs (SDPs). We consider large-scale POPs, for which interior-point methods are no longer able to solve the resulting SDPs. We propose an algorithm that combines a first-order method for solving the S… ▽ More The Moment/Sum-of-squares hierarchy provides a way to compute the global minimizers of polynomial optimization problems (POP), at the cost of solving a sequence of increasingly large semidefinite programs (SDPs). We consider large-scale POPs, for which interior-point methods are no longer able to solve the resulting SDPs. We propose an algorithm that combines a first-order method for solving the SDP relaxation, and a second-order method on a non-convex problem obtained from the POP. The switch from the first to the second-order method is based on a quantitative criterion, whose satisfaction ensures that Newton's method converges quadratically from its first iteration. This criterion leverages the point-estimation theory of Smale and the active-set identification. We illustrate the methodology to obtain global minimizers of large-scale optimal power flow problems. △ Less

Submitted 12 September, 2023; v1 submitted 25 May, 2023; originally announced May 2023.

MSC Class: 90C22; 90C23

Riemannian Stochastic Approximation for Minimizing Tame Nonsmooth Objective Functions

Authors: Johannes Aspman, Vyacheslav Kungurtsev, Reza Roohi Seraji

Abstract: In many learning applications, the parameters in a model are structurally constrained in a way that can be modeled as them lying on a Riemannian manifold. Riemannian optimization, wherein procedures to enforce an iterative minimizing sequence to be constrained to the manifold, is used to train such models. At the same time, tame geometry has become a significant topological description of nonsmoot… ▽ More In many learning applications, the parameters in a model are structurally constrained in a way that can be modeled as them lying on a Riemannian manifold. Riemannian optimization, wherein procedures to enforce an iterative minimizing sequence to be constrained to the manifold, is used to train such models. At the same time, tame geometry has become a significant topological description of nonsmooth functions that appear in the landscapes of training neural networks and other important models with structural compositions of continuous nonlinear functions with nonsmooth maps. In this paper, we study the properties of such stratifiable functions on a manifold and the behavior of retracted stochastic gradient descent, with diminishing stepsizes, for minimizing such functions. △ Less

Submitted 8 February, 2023; v1 submitted 1 February, 2023; originally announced February 2023.

Decay channels for double extremal black holes in four dimensions

Authors: Johannes Aspman, Jan Manschot

Abstract: We explore decay channels for charged black holes with vanishing temperature in $\mathcal{N}=2$ supersymmetric compactifications of string theory. If not protected by supersymmetry, such extremal black holes are expected to decay as a consequence of the weak gravity conjecture. We concentrate on double extremal, non-supersymmetric black holes for which the values of the scalar fields are constant… ▽ More We explore decay channels for charged black holes with vanishing temperature in $\mathcal{N}=2$ supersymmetric compactifications of string theory. If not protected by supersymmetry, such extremal black holes are expected to decay as a consequence of the weak gravity conjecture. We concentrate on double extremal, non-supersymmetric black holes for which the values of the scalar fields are constant throughout space-time, and explore decay channels for which decay into BPS and anti-BPS constituents is energetically favorable. We demonstrate the existence of decay channels at tree level for large families of double extremal black holes. For specific charges, we also find stable non-supersymmetric black holes, suggesting recombination of (anti)-supersymmetric constituents to a non-supersymmetric object. △ Less

Submitted 20 December, 2022; originally announced December 2022.

Comments: 3 figures

Journal ref: SciPost Phys. 16, 021 (2024)

Topological twists of massive SQCD, Part I

Authors: Johannes Aspman, Elias Furrer, Jan Manschot

Abstract: We consider topological twists of four-dimensional $\mathcal{N}=2$ supersymmetric QCD with gauge group SU(2) and $N_f\leq 3$ fundamental hypermultiplets. The twists are labelled by a choice of background fluxes for the flavour group, which provides an infinite family of topological partition functions. In this Part I, we demonstrate that in the presence of such fluxes the theories can be formulate… ▽ More We consider topological twists of four-dimensional $\mathcal{N}=2$ supersymmetric QCD with gauge group SU(2) and $N_f\leq 3$ fundamental hypermultiplets. The twists are labelled by a choice of background fluxes for the flavour group, which provides an infinite family of topological partition functions. In this Part I, we demonstrate that in the presence of such fluxes the theories can be formulated for arbitrary gauge bundles on a compact four-manifold. Moreover, we consider arbitrary masses for the hypermultiplets, which introduce new intricacies for the evaluation of the low-energy path integral on the Coulomb branch. We develop techniques for the evaluation of these path integrals. In the forthcoming Part II, we will deal with the explicit evaluation. △ Less

Submitted 20 December, 2023; v1 submitted 17 June, 2022; originally announced June 2022.

Comments: 41 pages + appendices, 8 figures. First part of a series of two papers, second part available at arXiv:2312.11616

Polynomial Matrix Inequalities within Tame Geometry

Authors: Christos Aravanis, Johannes Aspman, Georgios Korpas, Jakub Marecek

Abstract: Polynomial matrix inequalities can be solved using hierarchies of convex relaxations, pioneered by Henrion and Lassere. In some cases, this might not be practical, and one may need to resort to methods with local convergence guarantees, whose development has been rather ad hoc, so far. In this paper, we explore several alternative approaches to the problem, with non-trivial guarantees available us… ▽ More Polynomial matrix inequalities can be solved using hierarchies of convex relaxations, pioneered by Henrion and Lassere. In some cases, this might not be practical, and one may need to resort to methods with local convergence guarantees, whose development has been rather ad hoc, so far. In this paper, we explore several alternative approaches to the problem, with non-trivial guarantees available using results from tame geometry. △ Less

Submitted 8 June, 2022; originally announced June 2022.

Comments: 13 pages, 3 figures

Four flavours, triality and bimodular forms

Authors: Johannes Aspman, Elias Furrer, Jan Manschot

Abstract: We consider $\mathcal{N}=2$ supersymmetric $\text{SU}(2)$ gauge theory with $N_f=4$ massive hypermultiplets. The duality group of this theory contains transformations acting on the UV-coupling $τ_{\text{UV}}$ as well as on the running coupling $τ$. We establish that subgroups of the duality group act separately on $τ_{\text{UV}}$ and $τ$, while a larger group acts simultaneously on… ▽ More We consider $\mathcal{N}=2$ supersymmetric $\text{SU}(2)$ gauge theory with $N_f=4$ massive hypermultiplets. The duality group of this theory contains transformations acting on the UV-coupling $τ_{\text{UV}}$ as well as on the running coupling $τ$. We establish that subgroups of the duality group act separately on $τ_{\text{UV}}$ and $τ$, while a larger group acts simultaneously on $τ_{\text{UV}}$ and $τ$. For special choices of the masses, we find that the duality groups can be identified with congruence subgroups of $\text{SL}(2,\mathbb Z)$. We demonstrate that in such cases, the order parameters are instances of bimodular forms with arguments $τ$ and $τ_{\text{UV}}$. Since the UV duality group of the theory contains the triality group of outer automorphisms of the flavour symmetry $\text{SO}(8)$, the duality action gives rise to an orbit of mass configurations. Consequently, the corresponding order parameters combine to vector-valued bimodular forms with $\text{SL}(2,\mathbb Z)$ acting simultaneously on the two couplings. △ Less

Submitted 12 November, 2021; v1 submitted 22 October, 2021; originally announced October 2021.

Comments: 29 pages + Appendices, 7 figures, v2: minor changes, updated Appendix B

The $u$-plane integral, mock modularity and enumerative geometry

Authors: Johannes Aspman, Elias Furrer, Georgios Korpas, Zhi-Cong Ong, Meng-Chwan Tan

Abstract: We revisit the low-energy effective $U(1)$ action of topologically twisted $\mathcal N=2$ SYM theory with gauge group of rank one on a generic oriented smooth 4-manifold $X$ with nontrivial fundamental group. After including a specific new set of $\mathcal Q$-exact operators to the known action, we express the integrand of the path integral of the low-energy $U(1)$ theory as an anti-holomorphic de… ▽ More We revisit the low-energy effective $U(1)$ action of topologically twisted $\mathcal N=2$ SYM theory with gauge group of rank one on a generic oriented smooth 4-manifold $X$ with nontrivial fundamental group. After including a specific new set of $\mathcal Q$-exact operators to the known action, we express the integrand of the path integral of the low-energy $U(1)$ theory as an anti-holomorphic derivative. This allows us to use the theory of mock modular forms and indefinite theta functions for the explicit evaluation of correlation functions of the theory, including but not restricted to those that physically reproduce Donaldson invariants, thus facilitating the computations compared to previously used methods. As an explicit check of our results, we compute the path integral for the product ruled surfaces $X=Σ_g \times \mathbb{CP}^1$ for the reduction on either factor and compare the results with existing literature. In the case of reduction on the Riemann surface $Σ_g$, via an equivalent topological A-model on $\mathbb{CP}^1$, we will be able to express the generating function of genus zero Gromov-Witten invariants of the moduli space of flat rank one connections over $Σ_g$ in terms of an indefinite theta function, whence we would be able to make concrete numerical predictions of these enumerative invariants in terms of modular data, thereby allowing us to derive results in enumerative geometry from number theory. △ Less

Submitted 24 February, 2022; v1 submitted 9 September, 2021; originally announced September 2021.

Comments: 31 pages + appendices. Minor changes to title, abstract, introduction and section 5 after reviewer comments. To appear in Lett.Math.Phys

Journal ref: Lett Math Phys 112, 30 (2022)

Cutting and gluing with running couplings in $\mathcal{N}=2$ QCD

Authors: Johannes Aspman, Elias Furrer, Jan Manschot

Abstract: We consider the order parameter $u=\left<{\rm Tr}φ^2\right>$ as function of the running coupling constant $τ\in \mathbb{H}$ of asymptotically free $\mathcal{N}=2$ QCD with gauge group $SU(2)$ and $N_f\leq 3$ massive hypermultiplets. If the domain for $τ$ is restricted to an appropriate fundamental domain $\mathcal{F}_{N_f}$, the function $u$ is one-to-one. We demonstrate that these domains consist… ▽ More We consider the order parameter $u=\left<{\rm Tr}φ^2\right>$ as function of the running coupling constant $τ\in \mathbb{H}$ of asymptotically free $\mathcal{N}=2$ QCD with gauge group $SU(2)$ and $N_f\leq 3$ massive hypermultiplets. If the domain for $τ$ is restricted to an appropriate fundamental domain $\mathcal{F}_{N_f}$, the function $u$ is one-to-one. We demonstrate that these domains consist of six or less images of an ${\rm SL}(2,\mathbb{Z})$ keyhole fundamental domain, with appropriate identifications of the boundaries. For special choices of the masses, $u$ does not give rise to branch points and cuts, such that $u$ is a modular function for a congruence subgroup $Γ$ of ${\rm SL}(2,\mathbb{Z})$ and the fundamental domain is $Γ\backslash\mathbb{H}$. For generic masses, however, branch points and cuts are present, and subsets of $\mathcal{F}_{N_f}$ are being cut and glued upon varying the mass. We study this mechanism for various phenomena, such as decoupling of hypermultiplets, merging of local singularities, as well as merging of non-local singularities which give rise to superconformal Argyres-Douglas theories. △ Less

Submitted 9 July, 2021; originally announced July 2021.

Comments: 56 pages + Appendices, 22 figures

Elliptic Loci of SU(3) Vacua

Authors: Johannes Aspman, Elias Furrer, Jan Manschot

Abstract: The space of vacua of many four-dimensional, $\mathcal{N}=2$ supersymmetric gauge theories can famously be identified with a family of complex curves. For gauge group $SU(2)$, this gives a fully explicit description of the low-energy effective theory in terms of an elliptic curve and associated modular fundamental domain. The two-dimensional space of vacua for gauge group $SU(3)$ parametrizes an i… ▽ More The space of vacua of many four-dimensional, $\mathcal{N}=2$ supersymmetric gauge theories can famously be identified with a family of complex curves. For gauge group $SU(2)$, this gives a fully explicit description of the low-energy effective theory in terms of an elliptic curve and associated modular fundamental domain. The two-dimensional space of vacua for gauge group $SU(3)$ parametrizes an intricate family of genus two curves. We analyze this family using the so-called Rosenhain form for these curves. We demonstrate that two natural one-dimensional subloci of the space of $SU(3)$ vacua, $\mathcal{E}_u$ and $\mathcal{E}_v$, each parametrize a family of elliptic curves. For these elliptic loci, we describe the order parameters and fundamental domains explicitly. The locus $\mathcal{E}_u$ contains the points where mutually local dyons become massless, and is a fundamental domain for a classical congruence subgroup. Moreover, the locus $\mathcal{E}_v$ contains the superconformal Argyres-Douglas points, and is a fundamental domain for a Fricke group. △ Less

Submitted 17 May, 2021; v1 submitted 13 October, 2020; originally announced October 2020.

Comments: 39 pages + Appendices, 5 figures, v2: minor changes and extended discussion on automorphisms, v3: minor changes, published version

Report number: TCDMATH 20-09

Journal ref: Ann. Henri Poincaré (2021)