-
Speeding up Fermionic Lattice Calculations with Photonic Accelerated Inverters
Authors:
Felipe Attanasio,
Marc Bauer,
Jelle Dijkstra,
Timoteo Lee,
Jan M. Pawlowski,
Wolfram Pernice
Abstract:
Lattice field theory (LFT) is the standard non-perturbative method to perform numerical calculations of quantum field theory. However, the typical bottleneck of fermionic lattice calculations is the inversion of the Dirac matrix. This inversion is solved by iterative methods, like the conjugate gradient algorithm, where matrix-vector multiplications (MVMs) are the main operation. Photonic integrat…
▽ More
Lattice field theory (LFT) is the standard non-perturbative method to perform numerical calculations of quantum field theory. However, the typical bottleneck of fermionic lattice calculations is the inversion of the Dirac matrix. This inversion is solved by iterative methods, like the conjugate gradient algorithm, where matrix-vector multiplications (MVMs) are the main operation. Photonic integrated circuits excel in performing quick and energy-efficient MVMs, but at the same time, they are known to have low accuracy. This can be overcome by using mixed precision methods. In this paper, we explore the idea of using photonic technology to fulfil the demand for computational power of fermionic lattice calculations. These methods have the potential to reduce computation costs by one order of magnitude. Because of the hybrid nature of these methods, we call these 'photonic accelerated inverters (PAIs)'.
△ Less
Submitted 25 January, 2024;
originally announced January 2024.
-
QMeS-Derivation: Mathematica package for the symbolic derivation of functional equations
Authors:
Jan M. Pawlowski,
Coralie S. Schneider,
Nicolas Wink
Abstract:
We present the Mathematica package QMeS-Derivation. It derives symbolic functional equations from a given master equation. The latter include functional renormalisation group equations, Dyson-Schwinger equations, Slavnov-Taylor and Ward identities and their modifications in the presence of momentum cutoffs. The modules allow to derive the functional equations, take functional derivatives, trace ov…
▽ More
We present the Mathematica package QMeS-Derivation. It derives symbolic functional equations from a given master equation. The latter include functional renormalisation group equations, Dyson-Schwinger equations, Slavnov-Taylor and Ward identities and their modifications in the presence of momentum cutoffs. The modules allow to derive the functional equations, take functional derivatives, trace over field space, apply a given truncation scheme, and do momentum routings while kee** track of prefactors and signs that arise from fermionic commutation relations. The package furthermore contains an installer as well as Mathematica notebooks with showcase examples.
△ Less
Submitted 8 February, 2022; v1 submitted 2 February, 2021;
originally announced February 2021.
-
Towards Novel Insights in Lattice Field Theory with Explainable Machine Learning
Authors:
Stefan Bluecher,
Lukas Kades,
Jan M. Pawlowski,
Nils Strodthoff,
Julian M. Urban
Abstract:
Machine learning has the potential to aid our understanding of phase structures in lattice quantum field theories through the statistical analysis of Monte Carlo samples. Available algorithms, in particular those based on deep learning, often demonstrate remarkable performance in the search for previously unidentified features, but tend to lack transparency if applied naively. To address these sho…
▽ More
Machine learning has the potential to aid our understanding of phase structures in lattice quantum field theories through the statistical analysis of Monte Carlo samples. Available algorithms, in particular those based on deep learning, often demonstrate remarkable performance in the search for previously unidentified features, but tend to lack transparency if applied naively. To address these shortcomings, we propose representation learning in combination with interpretability methods as a framework for the identification of observables. More specifically, we investigate action parameter regression as a pretext task while using layer-wise relevance propagation (LRP) to identify the most important observables depending on the location in the phase diagram. The approach is put to work in the context of a scalar Yukawa model in (2+1)d. First, we investigate a multilayer perceptron to determine an importance hierarchy of several predefined, standard observables. The method is then applied directly to the raw field configurations using a convolutional network, demonstrating the ability to reconstruct all order parameters from the learned filter weights. Based on our results, we argue that due to its broad applicability, attribution methods such as LRP could prove a useful and versatile tool in our search for new physical insights. In the case of the Yukawa model, it facilitates the construction of an observable that characterises the symmetric phase.
△ Less
Submitted 18 May, 2020; v1 submitted 3 March, 2020;
originally announced March 2020.
-
Spectral Reconstruction with Deep Neural Networks
Authors:
Lukas Kades,
Jan M. Pawlowski,
Alexander Rothkopf,
Manuel Scherzer,
Julian M. Urban,
Sebastian J. Wetzel,
Nicolas Wink,
Felix P. G. Ziegler
Abstract:
We explore artificial neural networks as a tool for the reconstruction of spectral functions from imaginary time Green's functions, a classic ill-conditioned inverse problem. Our ansatz is based on a supervised learning framework in which prior knowledge is encoded in the training data and the inverse transformation manifold is explicitly parametrised through a neural network. We systematically in…
▽ More
We explore artificial neural networks as a tool for the reconstruction of spectral functions from imaginary time Green's functions, a classic ill-conditioned inverse problem. Our ansatz is based on a supervised learning framework in which prior knowledge is encoded in the training data and the inverse transformation manifold is explicitly parametrised through a neural network. We systematically investigate this novel reconstruction approach, providing a detailed analysis of its performance on physically motivated mock data, and compare it to established methods of Bayesian inference. The reconstruction accuracy is found to be at least comparable, and potentially superior in particular at larger noise levels. We argue that the use of labelled training data in a supervised setting and the freedom in defining an optimisation objective are inherent advantages of the present approach and may lead to significant improvements over state-of-the-art methods in the future. Potential directions for further research are discussed in detail.
△ Less
Submitted 1 February, 2021; v1 submitted 10 May, 2019;
originally announced May 2019.
-
Summary Report of Physics Beyond Colliders at CERN
Authors:
R. Alemany,
C. Burrage,
H. Bartosik,
J. Bernhard,
J. Boyd,
M. Brugger,
M. Calviani,
C. Carli,
N. Charitonidis,
D. Curtin,
A. Dainese,
A. de Roeck,
M. Diehl,
B. Döbrich,
L. Evans,
J. L. Feng,
M. Ferro-Luzzi,
L. Gatignon,
S. Gilardoni,
S. Gninenko,
G. Graziani,
E. Gschwendtner,
B. Goddard,
A. Hartin,
I. Irastorza
, et al. (39 additional authors not shown)
Abstract:
Physics Beyond Colliders is an exploratory study aimed at exploiting the full scientific potential of CERN's accelerator complex and its scientific infrastructure in the next two decades through projects complementary to the LHC, HL-LHC and other possible future colliders. These projects should target fundamental physics questions that are similar in spirit to those addressed by high-energy collid…
▽ More
Physics Beyond Colliders is an exploratory study aimed at exploiting the full scientific potential of CERN's accelerator complex and its scientific infrastructure in the next two decades through projects complementary to the LHC, HL-LHC and other possible future colliders. These projects should target fundamental physics questions that are similar in spirit to those addressed by high-energy colliders, but that require different types of beams and experiments. A kick-off workshop held in September 2016 identified a number of areas of interest and working groups have been set-up to study and develop these directions. All projects currently under consideration are presented including physics motivation, a brief outline of the experimental set-up and the status of the corresponding beam and detector technological studies. The proposals are also put in context of the worldwide landscape and their implementation issues are discussed.
△ Less
Submitted 1 February, 2019;
originally announced February 2019.
-
The Discrete Langevin Machine: Bridging the Gap Between Thermodynamic and Neuromorphic Systems
Authors:
Lukas Kades,
Jan M. Pawlowski
Abstract:
A formulation of Langevin dynamics for discrete systems is derived as a class of generic stochastic processes. The dynamics simplify for a two-state system and suggest a network architecture which is implemented by the Langevin machine. The Langevin machine represents a promising approach to compute successfully quantitative exact results of Boltzmann distributed systems by LIF neurons. Besides a…
▽ More
A formulation of Langevin dynamics for discrete systems is derived as a class of generic stochastic processes. The dynamics simplify for a two-state system and suggest a network architecture which is implemented by the Langevin machine. The Langevin machine represents a promising approach to compute successfully quantitative exact results of Boltzmann distributed systems by LIF neurons. Besides a detailed introduction of the dynamics, different simplified models of a neuromorphic hardware system are studied with respect to a control of emerging sources of errors.
△ Less
Submitted 7 April, 2021; v1 submitted 16 January, 2019;
originally announced January 2019.
-
Reducing Autocorrelation Times in Lattice Simulations with Generative Adversarial Networks
Authors:
Jan M. Pawlowski,
Julian M. Urban
Abstract:
Short autocorrelation times are essential for a reliable error assessment in Monte Carlo simulations of lattice systems. In many interesting scenarios, the decay of autocorrelations in the Markov chain is prohibitively slow. Generative samplers can provide statistically independent field configurations, thereby potentially ameliorating these issues. In this work, the applicability of neural sample…
▽ More
Short autocorrelation times are essential for a reliable error assessment in Monte Carlo simulations of lattice systems. In many interesting scenarios, the decay of autocorrelations in the Markov chain is prohibitively slow. Generative samplers can provide statistically independent field configurations, thereby potentially ameliorating these issues. In this work, the applicability of neural samplers to this problem is investigated. Specifically, we work with a generative adversarial network (GAN). We propose to address difficulties regarding its statistical exactness through the implementation of an overrelaxation step, by searching the latent space of the trained generator network. This procedure can be incorporated into a standard Monte Carlo algorithm, which then permits a sensible assessment of ergodicity and balance based on consistency checks. Numerical results for real, scalar $φ^4$-theory in two dimensions are presented. We achieve a significant reduction of autocorrelations while accurately reproducing the correct statistics. We discuss possible improvements to the approach as well as potential solutions to persisting issues.
△ Less
Submitted 23 December, 2021; v1 submitted 8 November, 2018;
originally announced November 2018.
-
Anomalous scaling at non-thermal fixed points of Burgers' and Gross-Pitaevskii turbulence
Authors:
Steven Mathey,
Thomas Gasenzer,
Jan M. Pawlowski
Abstract:
Scaling in the dynamical properties of complex many-body systems has been of strong interest since turbulence phenomena became the subject of systematic mathematical studies. In this article, dynamical critical phenomena far from equilibrium are investigated with functional renormalisation group equations. The focus is set on scaling solutions of the stochastic driven-dissipative Burgers equation…
▽ More
Scaling in the dynamical properties of complex many-body systems has been of strong interest since turbulence phenomena became the subject of systematic mathematical studies. In this article, dynamical critical phenomena far from equilibrium are investigated with functional renormalisation group equations. The focus is set on scaling solutions of the stochastic driven-dissipative Burgers equation and their relation to solutions known in the literature for Burgers and Kardar-Parisi-Zhang dynamics. We furthermore relate superfluid as well as acoustic turbulence described by the Gross-Pitaevskii model to known analytic and numerical results for scaling solutions. In this way, the canonical Kolmogorov exponent 5/3 for the energy cascade in superfluid turbulence is obtained analytically. We also get first results for anomalous exponents of acoustic and quantum turbulence. These are consistent with existing experimental data. Our results should be relevant for future experiments with, e.g., exciton-polariton condensates in solid-state systems as well as with ultra-cold atomic gases.
△ Less
Submitted 26 August, 2015; v1 submitted 29 May, 2014;
originally announced May 2014.
