Showing 1–20 of 20 results for author: Mbeng, G

Beyond Quantum Annealing: Optimal control solutions to MaxCut problems

Authors: Giovanni Pecci, Ruiyi Wang, Pietro Torta, Glen Bigan Mbeng, Giuseppe Santoro

Abstract: Quantum Annealing (QA) relies on mixing two Hamiltonian terms, a simple driver and a complex problem Hamiltonian, in a linear combination. The time-dependent schedule for this mixing is often taken to be linear in time: improving on this linear choice is known to be essential and has proven to be difficult. Here, we present different techniques for improving on the linear-schedule QA along two dir… ▽ More Quantum Annealing (QA) relies on mixing two Hamiltonian terms, a simple driver and a complex problem Hamiltonian, in a linear combination. The time-dependent schedule for this mixing is often taken to be linear in time: improving on this linear choice is known to be essential and has proven to be difficult. Here, we present different techniques for improving on the linear-schedule QA along two directions, conceptually distinct but leading to similar outcomes: 1) the first approach consists of constructing a Trotter-digitized QA (dQA) with schedules parameterized in terms of Fourier modes or Chebyshev polynomials, inspired by the Chopped Random Basis algorithm (CRAB) for optimal control in continuous time; 2) the second approach is technically a Quantum Approximate Optimization Algorithm (QAOA), whose solutions are found iteratively using linear interpolation or expansion in Fourier modes. Both approaches emphasize finding smooth optimal schedule parameters, ultimately leading to hybrid quantum-classical variational algorithms of the alternating Hamiltonian Ansatz type. We apply these techniques to MaxCut problems on weighted 3-regular graphs with N = 14 sites, focusing on hard instances that exhibit a small spectral gap, for which a standard linear-schedule QA performs poorly. We characterize the physics behind the optimal protocols for both the dQA and QAOA approaches, discovering shortcuts to adiabaticity-like dynamics. Furthermore, we study the transferability of such smooth solutions among hard instances of MaxCut at different circuit depths. Finally, we show that the smoothness pattern of these protocols obtained in a digital setting enables us to adapt them to continuous-time evolution, contrarily to generic non-smooth solutions. This procedure results in an optimized quantum annealing schedule that is implementable on analog devices. △ Less

Submitted 14 May, 2024; originally announced May 2024.

Comments: 18 pages, 13 figures

Error Mitigation for Quantum Approximate Optimization

Authors: Anita Weidinger, Glen Bigan Mbeng, Wolfgang Lechner

Abstract: Solving optimization problems on near term quantum devices requires develo** error mitigation techniques to cope with hardware decoherence and dephasing processes. We propose a mitigation technique based on the LHZ architecture. This architecture uses a redundant encoding of logical variables to solve optimization problems on fully programmable planar quantum chips. We discuss how this redundanc… ▽ More Solving optimization problems on near term quantum devices requires develo** error mitigation techniques to cope with hardware decoherence and dephasing processes. We propose a mitigation technique based on the LHZ architecture. This architecture uses a redundant encoding of logical variables to solve optimization problems on fully programmable planar quantum chips. We discuss how this redundancy can be exploited to mitigate errors in quantum optimization algorithms. In the specific context of the quantum approximate optimization algorithm (QAOA), we show that errors can be significantly mitigated by appropriately modifying the objective cost function. △ Less

Submitted 12 January, 2023; originally announced January 2023.

Rotated ansatz for approximate counterdiabatic driving

Authors: Glen Bigan Mbeng, Wolfgang Lechner

Abstract: Approximate counterdiabatic (CD) protocols are a powerful tool to enhance quantum adiabatic processes that allow to reliably manipulate quantum systems on short time scales. However, implementing CD protocols entails the introduction of additional control fields in the Hamiltonian, often associated with highly non-local multi-body interactions. Here, we introduce a novel variational rotated ansatz… ▽ More Approximate counterdiabatic (CD) protocols are a powerful tool to enhance quantum adiabatic processes that allow to reliably manipulate quantum systems on short time scales. However, implementing CD protocols entails the introduction of additional control fields in the Hamiltonian, often associated with highly non-local multi-body interactions. Here, we introduce a novel variational rotated ansatz (RA) to systematically generate experimentally accessible approximate CD protocols. We numerically benchmark our approach on state preparation and adiabatic quantum computing algorithms, and find that using RA protocols significantly enhances their performances. △ Less

Submitted 7 July, 2022; originally announced July 2022.

Symmetric projected entangled-pair states analysis of a phase transition in coupled spin-1/2 ladders

Authors: Juraj Hasik, Glen B. Mbeng, Sylvain Capponi, Federico Becca, Andreas M. Läuchli

Abstract: Infinite projected entangled-pair states (iPEPS) have been introduced to accurately describe many-body wave functions on two-dimensional lattices. In this context, two aspects are crucial: the systematic improvement of the {\it Ansatz} by the optimization of its building blocks, i.e., tensors characterized by bond dimension $D$, and the extrapolation scheme to reach the "thermodynamic" limit… ▽ More Infinite projected entangled-pair states (iPEPS) have been introduced to accurately describe many-body wave functions on two-dimensional lattices. In this context, two aspects are crucial: the systematic improvement of the {\it Ansatz} by the optimization of its building blocks, i.e., tensors characterized by bond dimension $D$, and the extrapolation scheme to reach the "thermodynamic" limit $D \to \infty$. Recent advances in variational optimization and scaling based on correlation lengths demonstrated the ability of iPEPS to capture the spontaneous breaking of a continuous symmetry in phases such as the antiferromagnetic (Néel) phase with high fidelity, in addition to valence-bond solids which are already well described by finite-$D$ iPEPS. In contrast, systems in the vicinity of continuous quantum phase transitions still present a challenge for iPEPS, especially when non-abelian symmetries are involved. Here, we consider the iPEPS Ansatz to describe the continuous transition between the (gapless) antiferromagnet and the (gapped) paramagnet that exists in the $S=1/2$ Heisenberg model on coupled two-leg ladders. In particular, we show how accurate iPEPS results can be obtained down to a narrow interval around criticality and analyze the scaling of the order parameter in the Néel phase in a spatially anisotropic situation. △ Less

Submitted 13 June, 2022; originally announced June 2022.

Comments: 9 pages, 7 figures, source code available at https://github.com/jurajHasik/peps-torch

Journal ref: Phys. Rev. B 106, 125154 (2022)

Avoiding barren plateaus via transferability of smooth solutions in Hamiltonian Variational Ansatz

Authors: Antonio Anna Mele, Glen Bigan Mbeng, Giuseppe Ernesto Santoro, Mario Collura, Pietro Torta

Abstract: A large ongoing research effort focuses on Variational Quantum Algorithms (VQAs), representing leading candidates to achieve computational speed-ups on current quantum devices. The scalability of VQAs to a large number of qubits, beyond the simulation capabilities of classical computers, is still debated. Two major hurdles are the proliferation of low-quality variational local minima, and the expo… ▽ More A large ongoing research effort focuses on Variational Quantum Algorithms (VQAs), representing leading candidates to achieve computational speed-ups on current quantum devices. The scalability of VQAs to a large number of qubits, beyond the simulation capabilities of classical computers, is still debated. Two major hurdles are the proliferation of low-quality variational local minima, and the exponential vanishing of gradients in the cost function landscape, a phenomenon referred to as barren plateaus. Here we show that by employing iterative search schemes one can effectively prepare the ground state of paradigmatic quantum many-body models, circumventing also the barren plateau phenomenon. This is accomplished by leveraging the transferability to larger system sizes of iterative solutions, displaying an intrinsic smoothness of the variational parameters, a result that does not extend to other solutions found via random-start local optimization. Our scheme could be directly tested on near-term quantum devices, running a refinement optimization in a favorable local landscape with non-vanishing gradients. △ Less

Submitted 2 January, 2023; v1 submitted 4 June, 2022; originally announced June 2022.

Comments: 5 + 9 pages, 4 + 12 figures, comments are welcome

Journal ref: Phys. Rev. A 106, L060401 (2021)

Two-dimensional $\mathbb{Z}_2$ lattice gauge theory on a near-term quantum simulator: variational quantum optimization, confinement, and topological order

Authors: Luca Lumia, Pietro Torta, Glen B. Mbeng, Giuseppe E. Santoro, Elisa Ercolessi, Michele Burrello, Matteo M. Wauters

Abstract: We propose an implementation of a two-dimensional $\mathbb{Z}_2$ lattice gauge theory model on a shallow quantum circuit, involving a number of single and two-qubits gates comparable to what can be achieved with present-day and near-future technologies. The ground state preparation is numerically analyzed on a small lattice with a variational quantum algorithm, which requires a small number of par… ▽ More We propose an implementation of a two-dimensional $\mathbb{Z}_2$ lattice gauge theory model on a shallow quantum circuit, involving a number of single and two-qubits gates comparable to what can be achieved with present-day and near-future technologies. The ground state preparation is numerically analyzed on a small lattice with a variational quantum algorithm, which requires a small number of parameters to reach high fidelities and can be efficiently scaled up on larger systems. Despite the reduced size of the lattice we consider, a transition between confined and deconfined regimes can be detected by measuring expectation values of Wilson loop operators or the topological entropy. Moreover, if periodic boundary conditions are implemented, the same optimal solution is transferable among all four different topological sectors, without any need for further optimization on the variational parameters. Our work shows that variational quantum algorithms provide a useful technique to be added in the growing toolbox for digital simulations of lattice gauge theories. △ Less

Submitted 5 July, 2022; v1 submitted 22 December, 2021; originally announced December 2021.

Comments: 22 pages, 19 figures

Report number: QDEV CMT NBI 2021

Quantum Approximate Optimization Algorithm applied to the binary perceptron

Authors: Pietro Torta, Glen B. Mbeng, Carlo Baldassi, Riccardo Zecchina, Giuseppe E. Santoro

Abstract: We apply digitized Quantum Annealing (QA) and Quantum Approximate Optimization Algorithm (QAOA) to a paradigmatic task of supervised learning in artificial neural networks: the optimization of synaptic weights for the binary perceptron. At variance with the usual QAOA applications to MaxCut, or to quantum spin-chains ground state preparation, the classical Hamiltonian is characterized by highly no… ▽ More We apply digitized Quantum Annealing (QA) and Quantum Approximate Optimization Algorithm (QAOA) to a paradigmatic task of supervised learning in artificial neural networks: the optimization of synaptic weights for the binary perceptron. At variance with the usual QAOA applications to MaxCut, or to quantum spin-chains ground state preparation, the classical Hamiltonian is characterized by highly non-local multi-spin interactions. Yet, we provide evidence for the existence of optimal smooth solutions for the QAOA parameters, which are transferable among typical instances of the same problem, and we prove numerically an enhanced performance of QAOA over traditional QA. We also investigate on the role of the QAOA optimization landscape geometry in this problem, showing that the detrimental effect of a gap-closing transition encountered in QA is also negatively affecting the performance of our implementation of QAOA. △ Less

Submitted 19 December, 2021; originally announced December 2021.

Comments: 14 pages, 9 figures

Polynomial scaling enhancement in ground-state preparation of Ising spin models via counter-diabatic driving

Authors: Andreas Hartmann, Glen Bigan Mbeng, Wolfgang Lechner

Abstract: The preparation of ground states of spin systems is a fundamental operation in quantum computing and serves as the basis of adiabatic quantum computing. This form of quantum computation is subject to the adiabatic theorem which in turn poses a fundamental speed limit. We show that by employing diabatic transitions via counter diabatic driving a less strict requirement on adiabaticity applies. We d… ▽ More The preparation of ground states of spin systems is a fundamental operation in quantum computing and serves as the basis of adiabatic quantum computing. This form of quantum computation is subject to the adiabatic theorem which in turn poses a fundamental speed limit. We show that by employing diabatic transitions via counter diabatic driving a less strict requirement on adiabaticity applies. We demonstrate a scaling advantage from local and multi-spin counter diabatic driving in the ground-state fidelity compared to their adiabatic counterpart, for different Ising spin models. △ Less

Submitted 10 September, 2021; originally announced September 2021.

Comments: 8 pages, 7 figures

Quantum optimization via four-body Rydberg gates

Authors: Clemens Dlaska, Kilian Ender, Glen Bigan Mbeng, Andreas Kruckenhauser, Wolfgang Lechner, Rick van Bijnen

Abstract: There is a large ongoing research effort towards obtaining a quantum advantage in the solution of combinatorial optimization problems on near-term quantum devices. A particularly promising platform for testing and develo** quantum optimization algorithms are arrays of trapped neutral atoms, laser-coupled to highly excited Rydberg states. However, encoding combinatorial optimization problems in a… ▽ More There is a large ongoing research effort towards obtaining a quantum advantage in the solution of combinatorial optimization problems on near-term quantum devices. A particularly promising platform for testing and develo** quantum optimization algorithms are arrays of trapped neutral atoms, laser-coupled to highly excited Rydberg states. However, encoding combinatorial optimization problems in atomic arrays is challenging due to the limited inter-qubit connectivity given by their native finite-range interactions. Here we propose and analyze a fast, high fidelity four-body Rydberg parity gate, enabling a direct and straightforward implementation of the Lechner-Hauke-Zoller (LHZ) scheme and its recent generalization, the parity architecture, a scalable architecture for encoding arbitrarily connected interaction graphs. Our gate relies on onetime-optimized adiabatic laser pulses and is fully programmable by adjusting two hold-times during operation. We numerically demonstrate an implementation of the quantum approximate optimization algorithm (QAOA) for a small scale test problem. Our approach allows for efficient execution of variational optimization steps with a constant number of system manipulations, independent of the system size, thus paving the way for experimental investigations of QAOA beyond the reach of numerical simulations. △ Less

Submitted 4 June, 2021; originally announced June 2021.

Comments: 11 pages

The quantum Ising chain for beginners

Authors: Glen Bigan Mbeng, Angelo Russomanno, Giuseppe E. Santoro

Abstract: We present here various techniques to work with clean and disordered quantum Ising chains, for the benefit of students and non-experts. Starting from the Jordan-Wigner transformation, which maps spin-1/2 systems into fermionic ones, we review some of the basic approaches to deal with the superconducting correlations that naturally emerge in this context. In particular, we analyse the form of the g… ▽ More We present here various techniques to work with clean and disordered quantum Ising chains, for the benefit of students and non-experts. Starting from the Jordan-Wigner transformation, which maps spin-1/2 systems into fermionic ones, we review some of the basic approaches to deal with the superconducting correlations that naturally emerge in this context. In particular, we analyse the form of the ground state and excitations of the model, relating them to the symmetry-breaking physics, and illustrate aspects connected to calculating dynamical quantities, thermal averages, correlation functions and entanglement entropy. A few problems provide simple applications of the techniques. △ Less

Submitted 19 June, 2024; v1 submitted 19 September, 2020; originally announced September 2020.

Journal ref: SciPost Phys. Lect. Notes 82 (2024)

Multi-spin counter-diabatic driving in many-body quantum Otto refrigerators

Authors: Andreas Hartmann, Victor Mukherjee, Glen Bigan Mbeng, Wolfgang Niedenzu, Wolfgang Lechner

Abstract: Quantum refrigerators pump heat from a cold to a hot reservoir. In the few-particle regime, counter-diabatic (CD) driving of, originally adiabatic, work-exchange strokes is a promising candidate to overcome the bottleneck of vanishing cooling power. Here, we present a finite-time many-body quantum refrigerator that yields finite cooling power at high coefficient of performance, that considerably o… ▽ More Quantum refrigerators pump heat from a cold to a hot reservoir. In the few-particle regime, counter-diabatic (CD) driving of, originally adiabatic, work-exchange strokes is a promising candidate to overcome the bottleneck of vanishing cooling power. Here, we present a finite-time many-body quantum refrigerator that yields finite cooling power at high coefficient of performance, that considerably outperforms its non-adiabatic counterpart. We employ multi-spin CD driving and numerically investigate the scaling behavior of the refrigeration performance with system size. We further prove that optimal refrigeration via the exact CD protocol is a catalytic process. △ Less

Submitted 21 December, 2020; v1 submitted 21 August, 2020; originally announced August 2020.

Comments: 13 pages, 5 figures

Journal ref: Quantum 4, 377 (2020)

Reinforcement Learning assisted Quantum Optimization

Authors: Matteo M. Wauters, Emanuele Panizon, Glen B. Mbeng, Giuseppe E. Santoro

Abstract: We propose a reinforcement learning (RL) scheme for feedback quantum control within the quan-tum approximate optimization algorithm (QAOA). QAOA requires a variational minimization for states constructed by applying a sequence of unitary operators, depending on parameters living ina highly dimensional space. We reformulate such a minimum search as a learning task, where a RL agent chooses the cont… ▽ More We propose a reinforcement learning (RL) scheme for feedback quantum control within the quan-tum approximate optimization algorithm (QAOA). QAOA requires a variational minimization for states constructed by applying a sequence of unitary operators, depending on parameters living ina highly dimensional space. We reformulate such a minimum search as a learning task, where a RL agent chooses the control parameters for the unitaries, given partial information on the system. We show that our RL scheme finds a policy converging to the optimal adiabatic solution for QAOA found by Mbeng et al. arXiv:1906.08948 for the translationally invariant quantum Ising chain. In presence of disorder, we show that our RL scheme allows the training part to be performed on small samples, and transferred successfully on larger systems. △ Less

Submitted 26 April, 2020; originally announced April 2020.

Comments: 4 pages, 3 figures

Journal ref: Phys. Rev. Research 2, 033446 (2020)

Polynomial scaling of QAOA for ground-state preparation of the fully-connected p-spin ferromagnet

Authors: Matteo M. Wauters, Glen Bigan Mbeng, Giuseppe E. Santoro

Abstract: We show that the quantum approximate optimization algorithm (QAOA) can construct with polynomially scaling resources the ground state of the fully-connected p-spin Ising ferromagnet, a problem that notoriously poses severe difficulties to a Quantum Annealing (QA) approach, due to the exponentially small gaps encountered at first-order phase transition for ${\rm p} \ge 3$. For a target ground state… ▽ More We show that the quantum approximate optimization algorithm (QAOA) can construct with polynomially scaling resources the ground state of the fully-connected p-spin Ising ferromagnet, a problem that notoriously poses severe difficulties to a Quantum Annealing (QA) approach, due to the exponentially small gaps encountered at first-order phase transition for ${\rm p} \ge 3$. For a target ground state at arbitrary transverse field, we find that an appropriate QAOA parameter initialization is necessary to achieve a good performance of the algorithm when the number of variational parameters $2{\rm P}$ is much smaller than the system size ${\rm N}$, because of the large number of sub-optimal local minima. Instead, when ${\rm P}$ exceeds a critical value ${\rm P}^*_{\rm N} \propto {\rm N}$, the structure of the parameter space simplifies, as all minima become degenerate. This allows to achieve the ground state with perfect fidelity with a number of parameters scaling extensively with ${\rm N}$, and with resources scaling polynomially with ${\rm N}$. △ Less

Submitted 5 August, 2020; v1 submitted 16 March, 2020; originally announced March 2020.

Comments: 4 pages, 4 figures, Appendices 5 pages

Optimal quantum control with digitized Quantum Annealing

Authors: Glen Bigan Mbeng, Rosario Fazio, Giuseppe E. Santoro

Abstract: We show how a digitized version of Quantum Annealing can be made optimal, realizing the best possible solution allowed by quantum mechanics in the shortest time, without any prior knowledge on the location and properties of the spectral gap. Our findings elucidate the intimate relation between digitized-QA, optimal Quantum Control, and recently proposed hybrid quantum-classical variational algorit… ▽ More We show how a digitized version of Quantum Annealing can be made optimal, realizing the best possible solution allowed by quantum mechanics in the shortest time, without any prior knowledge on the location and properties of the spectral gap. Our findings elucidate the intimate relation between digitized-QA, optimal Quantum Control, and recently proposed hybrid quantum-classical variational algorithms for quantum-state preparation and optimization. We illustrate this on the simple benchmark problem of an unfrustrated antiferromagnetic Ising chain in a transverse field. △ Less

Submitted 25 November, 2019; originally announced November 2019.

Comments: arXiv admin note: substantial text overlap with arXiv:1906.08948

ACM Class: J.2

Optimal working point in digitized quantum annealing

Authors: Glen Bigan Mbeng, Luca Arceci, Giuseppe E. Santoro

Abstract: We present a study of the digitized Quantum Annealing protocol proposed by R. Barends et al., Nature 534, 222 (2016). Our analysis, performed on the benchmark case of a transverse Ising chain problem, shows that the algorithm has a well defined optimal working point for the annealing time $τ^{\mathrm{opt}}_\mathrm{P}$ --- scaling as $τ^{\mathrm{opt}}_\mathrm{P}\sim \mathrm{P}$, where $\mathrm{P}$… ▽ More We present a study of the digitized Quantum Annealing protocol proposed by R. Barends et al., Nature 534, 222 (2016). Our analysis, performed on the benchmark case of a transverse Ising chain problem, shows that the algorithm has a well defined optimal working point for the annealing time $τ^{\mathrm{opt}}_\mathrm{P}$ --- scaling as $τ^{\mathrm{opt}}_\mathrm{P}\sim \mathrm{P}$, where $\mathrm{P}$ is the number of digital Trotter steps --- beyond which, the residual energy error shoots-up towards the value characteristic of the maximally disordered state. We present an analytical analysis for the translationally invariant transverse Ising chain case, but our numerical evidence suggests that this scenario is more general, surviving, for instance, the presence of disorder. △ Less

Submitted 2 September, 2019; originally announced September 2019.

Journal ref: Phys. Rev. B 100, 224201 (2019)

Quantum Annealing: a journey through Digitalization, Control, and hybrid Quantum Variational schemes

Authors: Glen Bigan Mbeng, Rosario Fazio, Giuseppe Santoro

Abstract: We establish and discuss a number of connections between a digitized version of Quantum Annealing (QA) with the Quantum Approximate Optimization Algorithm (QAOA) introduced by Farhi et al. (arXiv:1411.4028) as an alternative hybrid quantum-classical variational scheme for quantum-state preparation and optimization. We introduce a technique that allows to prove, for instance, a rigorous bound conce… ▽ More We establish and discuss a number of connections between a digitized version of Quantum Annealing (QA) with the Quantum Approximate Optimization Algorithm (QAOA) introduced by Farhi et al. (arXiv:1411.4028) as an alternative hybrid quantum-classical variational scheme for quantum-state preparation and optimization. We introduce a technique that allows to prove, for instance, a rigorous bound concerning the performance of QAOA for MaxCut on a $2$-regular graph, equivalent to an unfrustrated antiferromagnetic Ising chain. The bound shows that the optimal variational error of a depth-$\mathrm{P}$ quantum circuit has to satisfy $ε^\mathrm{res}_{\mathrm{P}}\ge (2\mathrm{P}+2)^{-1}$. In a separate work (Mbeng et al., arXiv:1911.12259) we have explicitly shown, exploiting a Jordan-Wigner transformation, that among the $2^{\mathrm{P}}$ degenerate variational minima which can be found for this problem, all strictly satisfying the equality $ε^\mathrm{res}_{\mathrm{P}}=(2\mathrm{P}+2)^{-1}$, one can construct a special {\em regular} optimal solution, which is computationally optimal and does not require any prior knowledge about the spectral gap. We explicitly demonstrate here that such a schedule is adiabatic, in a digitized sense, and can therefore be interpreted as an optimized digitized-QA protocol. We also discuss and compare our bound on the residual energy to well-known results on the Kibble-Zurek mechanism behind a continuous-time QA. These findings help elucidating the intimate relation between digitized-QA, QAOA, and optimal Quantum Control. △ Less

Submitted 9 December, 2019; v1 submitted 21 June, 2019; originally announced June 2019.

MSC Class: J.2 ACM Class: J.2

The Information Complexity of Learning Tasks, their Structure and their Distance

Authors: Alessandro Achille, Giovanni Paolini, Glen Mbeng, Stefano Soatto

Abstract: We introduce an asymmetric distance in the space of learning tasks, and a framework to compute their complexity. These concepts are foundational for the practice of transfer learning, whereby a parametric model is pre-trained for a task, and then fine-tuned for another. The framework we develop is non-asymptotic, captures the finite nature of the training dataset, and allows distinguishing learnin… ▽ More We introduce an asymmetric distance in the space of learning tasks, and a framework to compute their complexity. These concepts are foundational for the practice of transfer learning, whereby a parametric model is pre-trained for a task, and then fine-tuned for another. The framework we develop is non-asymptotic, captures the finite nature of the training dataset, and allows distinguishing learning from memorization. It encompasses, as special cases, classical notions from Kolmogorov complexity, Shannon, and Fisher Information. However, unlike some of those frameworks, it can be applied to large-scale models and real-world datasets. Our framework is the first to measure complexity in a way that accounts for the effect of the optimization scheme, which is critical in Deep Learning. △ Less

Submitted 14 July, 2020; v1 submitted 5 April, 2019; originally announced April 2019.

Report number: UCLA CSD180003

Dynamics and Reachability of Learning Tasks

Authors: Alessandro Achille, Glen Mbeng, Stefano Soatto

Abstract: We compute the transition probability between two learning tasks, and show that it decomposes into two factors. The first depends on the geometry of the loss landscape of a model trained on each task, independent of any particular model used. This is related to an information theoretic distance function, but is insufficient to predict success in transfer learning, as nearby tasks can be unreachabl… ▽ More We compute the transition probability between two learning tasks, and show that it decomposes into two factors. The first depends on the geometry of the loss landscape of a model trained on each task, independent of any particular model used. This is related to an information theoretic distance function, but is insufficient to predict success in transfer learning, as nearby tasks can be unreachable via fine-tuning. The second factor depends on the ease of traversing the path between two tasks. With this dynamic component, we derive strict lower bounds on the complexity necessary to learn a task starting from the solution to another, which is one of the most common forms of transfer learning. △ Less

Submitted 29 May, 2019; v1 submitted 4 October, 2018; originally announced October 2018.

On the dynamics of Simulated Quantum Annealing in random Ising chains

Authors: Glen Bigan Mbeng, Lorenzo Privitera, Luca Arceci, Giuseppe E. Santoro

Abstract: Simulated Quantum Annealing (SQA), that is emulating a Quantum Annealing (QA) dynamics on a classical computer by a Quantum Monte Carlo whose parameters are changed during the simulation, is a well established computational strategy to cope with the exponentially large Hilbert space. It has enjoyed some early successes but has also raised more recent criticisms. Here we investigate, on the paradig… ▽ More Simulated Quantum Annealing (SQA), that is emulating a Quantum Annealing (QA) dynamics on a classical computer by a Quantum Monte Carlo whose parameters are changed during the simulation, is a well established computational strategy to cope with the exponentially large Hilbert space. It has enjoyed some early successes but has also raised more recent criticisms. Here we investigate, on the paradigmatic case of a one-dimensional transverse field Ising chain, two issues related to SQA in its Path-Integral implementation: the question of Monte Carlo vs physical (Schrödinger) dynamics and the issue of the imaginary-time continuum limit to eliminate the Trotter error. We show that, while a proper time-continuum limit is able to restitute the correct Kibble-Zurek scaling of the residual energy $\varepsilon_\mathrm{res}(τ)\sim τ^{-1/2}$ for the ordered case --- $τ$ being the total annealing time ---, the presence of disorder leads to a characteristic sampling crisis for a large number of Trotter time-slices, in the low-temperature ordered phase. Such sampling problem, in turn, leads to SQA results which are apparently unrelated to the coherent Schrödinger QA even at intermediate $τ$. △ Less

Submitted 6 September, 2018; originally announced September 2018.

Comments: 8 pages, 8 figures

Journal ref: Phys. Rev. B 99, 064201 (2019)

Negativity spectrum in 1D gapped phases of matter

Authors: Glen Bigan Mbeng, Vincenzo Alba, Pasquale Calabrese

Abstract: We investigate the spectrum of the partial transpose (negativity spectrum) of two adjacent regions in gapped one-dimensional models. We show that, in the limit of large regions, the negativity spectrum is entirely reconstructed from the entanglement spectrum of the bipartite system. We exploit this result in the XXZ spin chain, for which the entanglement spectrum is known by means of the corner tr… ▽ More We investigate the spectrum of the partial transpose (negativity spectrum) of two adjacent regions in gapped one-dimensional models. We show that, in the limit of large regions, the negativity spectrum is entirely reconstructed from the entanglement spectrum of the bipartite system. We exploit this result in the XXZ spin chain, for which the entanglement spectrum is known by means of the corner transfer matrix. We find that the negativity spectrum levels are equally spaced, the spacing being half that in the entanglement spectrum. Moreover, the degeneracy of the spectrum is described by elegant combinatorial formulas, which are related to the counting of integer partitions. We also derive the asymptotic distribution of the negativity spectrum. We provide exact results for the logarithmic negativity and for the moments of the partial transpose. They exhibit unusual scaling corrections in the limit $Δ\to1^+$ with a corrections exponent which is the same as that for the Rényi entropies. △ Less

Submitted 15 April, 2017; v1 submitted 15 December, 2016; originally announced December 2016.

Comments: 20 pages, 7 figures. As published. Contribution to the special issue of J. Phys. A: "John Cardy's scale-invariant journey in low dimensions: a special issue for his 70th birthday"

Journal ref: J. Phys. A: Math. Theor. 50, 194001 (2017)