-
Understanding multi-fidelity training of machine-learned force-fields
Authors:
John L. A. Gardner,
Hannes Schulz,
Jean Helie,
Lixin Sun,
Gregor N. C. Simm
Abstract:
Effectively leveraging data from multiple quantum-chemical methods is essential for building machine-learned force fields (MLFFs) that are applicable to a wide range of chemical systems. This study systematically investigates two multi-fidelity training strategies, pre-training/fine-tuning and multi-headed training, to elucidate the mechanisms underpinning their success. We identify key factors dr…
▽ More
Effectively leveraging data from multiple quantum-chemical methods is essential for building machine-learned force fields (MLFFs) that are applicable to a wide range of chemical systems. This study systematically investigates two multi-fidelity training strategies, pre-training/fine-tuning and multi-headed training, to elucidate the mechanisms underpinning their success. We identify key factors driving the efficacy of pre-training followed by fine-tuning, but find that internal representations learned during pre-training are inherently method-specific, requiring adaptation of the model backbone during fine-tuning. Multi-headed models offer an extensible alternative, enabling simultaneous training on multiple fidelities. We demonstrate that a multi-headed model learns method-agnostic representations that allow for accurate predictions across multiple label sources. While this approach introduces a slight accuracy compromise compared to sequential fine-tuning, it unlocks new cost-efficient data generation strategies and paves the way towards developing universal MLFFs.
△ Less
Submitted 17 June, 2025;
originally announced June 2025.
-
Asymptotics for a class of planar orthogonal polynomials and truncated unitary matrices
Authors:
Alfredo Deaño,
Kenneth T-R McLaughlin,
Leslie Molag,
Nick Simm
Abstract:
We carry out the asymptotic analysis as $n \to \infty$ of a class of orthogonal polynomials $p_{n}(z)$ of degree $n$, defined with respect to the planar measure \begin{equation*} dμ(z) = (1-|z|^{2})^{α-1}|z-x|^γ\mathbf{1}_{|z| < 1}d^{2}z, \end{equation*} where $d^{2}z$ is the two dimensional area measure, $α$ is a parameter that can grow with $n$, while $γ>-2$ and $x>0$ are fixed. This measure ari…
▽ More
We carry out the asymptotic analysis as $n \to \infty$ of a class of orthogonal polynomials $p_{n}(z)$ of degree $n$, defined with respect to the planar measure \begin{equation*} dμ(z) = (1-|z|^{2})^{α-1}|z-x|^γ\mathbf{1}_{|z| < 1}d^{2}z, \end{equation*} where $d^{2}z$ is the two dimensional area measure, $α$ is a parameter that can grow with $n$, while $γ>-2$ and $x>0$ are fixed. This measure arises naturally in the study of characteristic polynomials of non-Hermitian ensembles and generalises the example of a Gaussian weight that was recently studied by several authors. We obtain asymptotics in all regions of the complex plane and via an appropriate differential identity, we obtain the asymptotic expansion of the partition function. The main approach is to convert the planar orthogonality to one defined on suitable contours in the complex plane. Then the asymptotic analysis is performed using the Deift-Zhou steepest descent method for the associated Riemann-Hilbert problem.
△ Less
Submitted 5 June, 2025; v1 submitted 18 May, 2025;
originally announced May 2025.
-
The Fourier coefficients of the holomorphic multiplicative chaos in the limit of large frequency
Authors:
Joseph Najnudel,
Elliot Paquette,
Nick Simm,
Truong Vu
Abstract:
The holomorphic multiplicative chaos (HMC) is a holomorphic analogue of the Gaussian multiplicative chaos. It arises naturally as the limit in large matrix size of the characteristic polynomial of Haar unitary, and more generally circular-$β$-ensemble, random matrices.
We consider the Fourier coefficients of the holomorphic multiplicative chaos in the $L^1$-phase, and we show that appropriately…
▽ More
The holomorphic multiplicative chaos (HMC) is a holomorphic analogue of the Gaussian multiplicative chaos. It arises naturally as the limit in large matrix size of the characteristic polynomial of Haar unitary, and more generally circular-$β$-ensemble, random matrices.
We consider the Fourier coefficients of the holomorphic multiplicative chaos in the $L^1$-phase, and we show that appropriately normalized, this converges in distribution to a complex normal random variable, scaled by the total mass of the Gaussian multiplicative chaos measure on the unit circle. We further generalize this to a process convergence, showing the joint convergence of consecutive Fourier coefficients. As a corollary, we derive convergence in law of the secular coefficients of sublinear index of the circular-$β$-ensemble for all $β> 2$.
△ Less
Submitted 20 February, 2025;
originally announced February 2025.
-
On moments of the derivative of CUE characteristic polynomials and the Riemann zeta function
Authors:
Nick Simm,
Fei Wei
Abstract:
We study the derivative of the characteristic polynomial of $N \times N$ Haar distributed unitary matrices. We obtain the first explicit formulae for complex-valued moments when the spectral variable is inside the unit disc, in the limit $N \to \infty$. These formulae are expressed in terms of the confluent hypergeometric function of the first kind. As an application, we provide an alternative met…
▽ More
We study the derivative of the characteristic polynomial of $N \times N$ Haar distributed unitary matrices. We obtain the first explicit formulae for complex-valued moments when the spectral variable is inside the unit disc, in the limit $N \to \infty$. These formulae are expressed in terms of the confluent hypergeometric function of the first kind. As an application, we provide an alternative method to re-obtain Mezzadri's result [J. Phys. A, 36(12):2945-2962, 2003] on the asymptotic density of zeros of the derivative as $N \to \infty$. We explore the connection between these moments and those of the derivative of the Riemann zeta function away from the critical line. Under the Lindelöf hypothesis, we prove that all positive integer moments agree with our random matrix results up to an arithmetic factor. Inspired by this finding, we propose a conjecture on the asymptotics of non-integer moments of the derivative of the Riemann zeta function off the critical line. Within random matrix theory, we also investigate the microscopic regime where the spectral variable $z$ satisfies $|z|^{2}=1-\frac{c}{N}$ for a fixed constant $c$. We obtain an asymptotic formula for the moments in this regime as a determinant involving the finite temperature Bessel kernel, which reduces to the Bessel kernel when $c=0$. For finite matrix size, we provide an exact formula for the moments of the derivative inside the unit disc, expressed as polynomials of the inverse of the distance from the circle, with coefficients given by combinatorial sums.
△ Less
Submitted 23 December, 2024; v1 submitted 5 September, 2024;
originally announced September 2024.
-
Schur function expansion in non-Hermitian ensembles and averages of characteristic polynomials
Authors:
Alexander Serebryakov,
Nick Simm
Abstract:
We study $k$-point correlators of characteristic polynomials in non-Hermitian ensembles of random matrices, focusing on the real, complex and quaternion $N \times N$ Ginibre ensembles. Our approach is based on the technique of character expansions, which expresses the correlator as a sum over partitions involving Schur functions. We show how to re-sum the expansions in terms of representations whi…
▽ More
We study $k$-point correlators of characteristic polynomials in non-Hermitian ensembles of random matrices, focusing on the real, complex and quaternion $N \times N$ Ginibre ensembles. Our approach is based on the technique of character expansions, which expresses the correlator as a sum over partitions involving Schur functions. We show how to re-sum the expansions in terms of representations which interchange the roles of $N$ and $k$. We also provide a probabilistic interpretation of the character expansion analogous to the Schur measure, linking the correlators to the distribution of the top row in certain Young diagrams. In more specific examples we evaluate these expressions explicitly in terms of $k \times k$ determinants or Pfaffians. We show that our approach extends to other ensembles, such as truncations of random unitary matrices.
△ Less
Submitted 12 July, 2024; v1 submitted 31 October, 2023;
originally announced October 2023.
-
Large deviations and fluctuations of real eigenvalues of elliptic random matrices
Authors:
Sung-Soo Byun,
Leslie Molag,
Nick Simm
Abstract:
We study real eigenvalues of $N\times N$ real elliptic Ginibre matrices indexed by a non-Hermiticity parameter $0\leq τ<1$, in both the strong and weak non-Hermiticity regime. Here $N$ is assumed to be an even number. In both regimes, we prove a central limit theorem for the number of real eigenvalues. We also find the asymptotic behaviour of the probability $p_{N,k}^{(τ)}$ that exactly $k$ eigenv…
▽ More
We study real eigenvalues of $N\times N$ real elliptic Ginibre matrices indexed by a non-Hermiticity parameter $0\leq τ<1$, in both the strong and weak non-Hermiticity regime. Here $N$ is assumed to be an even number. In both regimes, we prove a central limit theorem for the number of real eigenvalues. We also find the asymptotic behaviour of the probability $p_{N,k}^{(τ)}$ that exactly $k$ eigenvalues are real. In the strong non-Hermiticity regime, where $τ$ is fixed, we find \begin{align*}
\lim_{N\to\infty} \frac{1}{\sqrt{N}} \log p_{N,k_N}^{(τ)} = -\sqrt\frac{1+τ}{1-τ} \frac{ζ(3/2)}{\sqrt{2π}} \end{align*} for any sequence $(k_N)_N$ of even numbers such that $k_N = o(\frac{\sqrt N}{\log N})$ as $N\to\infty$, where $ζ$ is the Riemann zeta function. In the weak non-Hermiticity regime, where $τ=1-\frac{α^2}{N}$, we obtain \begin{align*}
\lim_{N\to\infty} \frac{1}{N} \log p_{N,k_N}^{(τ)} \leq \frac{2}π \int_0^1 \log\left(1-e^{-α^2 s^2}\right) \sqrt{1-s^2} \, ds \end{align*} for any sequence $(k_N)_N$ of even numbers such that $k_N=o(\frac{N}{\log N})$ as $n\to\infty$. This inequality is expected to be an equality.
△ Less
Submitted 4 May, 2023;
originally announced May 2023.
-
Flow Annealed Importance Sampling Bootstrap
Authors:
Laurence Illing Midgley,
Vincent Stimper,
Gregor N. C. Simm,
Bernhard Schölkopf,
José Miguel Hernández-Lobato
Abstract:
Normalizing flows are tractable density models that can approximate complicated target distributions, e.g. Boltzmann distributions of physical systems. However, current methods for training flows either suffer from mode-seeking behavior, use samples from the target generated beforehand by expensive MCMC methods, or use stochastic losses that have high variance. To avoid these problems, we augment…
▽ More
Normalizing flows are tractable density models that can approximate complicated target distributions, e.g. Boltzmann distributions of physical systems. However, current methods for training flows either suffer from mode-seeking behavior, use samples from the target generated beforehand by expensive MCMC methods, or use stochastic losses that have high variance. To avoid these problems, we augment flows with annealed importance sampling (AIS) and minimize the mass-covering $α$-divergence with $α=2$, which minimizes importance weight variance. Our method, Flow AIS Bootstrap (FAB), uses AIS to generate samples in regions where the flow is a poor approximation of the target, facilitating the discovery of new modes. We apply FAB to multimodal targets and show that we can approximate them very accurately where previous methods fail. To the best of our knowledge, we are the first to learn the Boltzmann distribution of the alanine dipeptide molecule using only the unnormalized target density, without access to samples generated via Molecular Dynamics (MD) simulations: FAB produces better results than training via maximum likelihood on MD samples while using 100 times fewer target evaluations. After reweighting the samples, we obtain unbiased histograms of dihedral angles that are almost identical to the ground truth.
△ Less
Submitted 7 March, 2023; v1 submitted 3 August, 2022;
originally announced August 2022.
-
MACE: Higher Order Equivariant Message Passing Neural Networks for Fast and Accurate Force Fields
Authors:
Ilyes Batatia,
Dávid Péter Kovács,
Gregor N. C. Simm,
Christoph Ortner,
Gábor Csányi
Abstract:
Creating fast and accurate force fields is a long-standing challenge in computational chemistry and materials science. Recently, several equivariant message passing neural networks (MPNNs) have been shown to outperform models built using other approaches in terms of accuracy. However, most MPNNs suffer from high computational cost and poor scalability. We propose that these limitations arise becau…
▽ More
Creating fast and accurate force fields is a long-standing challenge in computational chemistry and materials science. Recently, several equivariant message passing neural networks (MPNNs) have been shown to outperform models built using other approaches in terms of accuracy. However, most MPNNs suffer from high computational cost and poor scalability. We propose that these limitations arise because MPNNs only pass two-body messages leading to a direct relationship between the number of layers and the expressivity of the network. In this work, we introduce MACE, a new equivariant MPNN model that uses higher body order messages. In particular, we show that using four-body messages reduces the required number of message passing iterations to just two, resulting in a fast and highly parallelizable model, reaching or exceeding state-of-the-art accuracy on the rMD17, 3BPA, and AcAc benchmark tasks. We also demonstrate that using higher order messages leads to an improved steepness of the learning curves.
△ Less
Submitted 26 January, 2023; v1 submitted 15 June, 2022;
originally announced June 2022.
-
The Design Space of E(3)-Equivariant Atom-Centered Interatomic Potentials
Authors:
Ilyes Batatia,
Simon Batzner,
Dávid Péter Kovács,
Albert Musaelian,
Gregor N. C. Simm,
Ralf Drautz,
Christoph Ortner,
Boris Kozinsky,
Gábor Csányi
Abstract:
The rapid progress of machine learning interatomic potentials over the past couple of years produced a number of new architectures. Particularly notable among these are the Atomic Cluster Expansion (ACE), which unified many of the earlier ideas around atom density-based descriptors, and Neural Equivariant Interatomic Potentials (NequIP), a message passing neural network with equivariant features t…
▽ More
The rapid progress of machine learning interatomic potentials over the past couple of years produced a number of new architectures. Particularly notable among these are the Atomic Cluster Expansion (ACE), which unified many of the earlier ideas around atom density-based descriptors, and Neural Equivariant Interatomic Potentials (NequIP), a message passing neural network with equivariant features that showed state of the art accuracy. In this work, we construct a mathematical framework that unifies these models: ACE is generalised so that it can be recast as one layer of a multi-layer architecture. From another point of view, the linearised version of NequIP is understood as a particular sparsification of a much larger polynomial model. Our framework also provides a practical tool for systematically probing different choices in the unified design space. We demonstrate this by an ablation study of NequIP via a set of experiments looking at in- and out-of-domain accuracy and smooth extrapolation very far from the training data, and shed some light on which design choices are critical for achieving high accuracy. Finally, we present BOTNet (Body-Ordered-Tensor-Network), a much-simplified version of NequIP, which has an interpretable architecture and maintains accuracy on benchmark datasets.
△ Less
Submitted 24 November, 2022; v1 submitted 13 May, 2022;
originally announced May 2022.
-
Bootstrap Your Flow
Authors:
Laurence Illing Midgley,
Vincent Stimper,
Gregor N. C. Simm,
José Miguel Hernández-Lobato
Abstract:
Normalizing flows are flexible, parameterized distributions that can be used to approximate expectations from intractable distributions via importance sampling. However, current flow-based approaches are limited on challenging targets where they either suffer from mode seeking behaviour or high variance in the training loss, or rely on samples from the target distribution, which may not be availab…
▽ More
Normalizing flows are flexible, parameterized distributions that can be used to approximate expectations from intractable distributions via importance sampling. However, current flow-based approaches are limited on challenging targets where they either suffer from mode seeking behaviour or high variance in the training loss, or rely on samples from the target distribution, which may not be available. To address these challenges, we combine flows with annealed importance sampling (AIS), while using the $α$-divergence as our objective, in a novel training procedure, FAB (Flow AIS Bootstrap). Thereby, the flow and AIS improve each other in a bootstrapping manner. We demonstrate that FAB can be used to produce accurate approximations to complex target distributions, including Boltzmann distributions, in problems where previous flow-based methods fail.
△ Less
Submitted 14 March, 2022; v1 submitted 22 November, 2021;
originally announced November 2021.
-
DOCKSTRING: easy molecular docking yields better benchmarks for ligand design
Authors:
Miguel García-Ortegón,
Gregor N. C. Simm,
Austin J. Tripp,
José Miguel Hernández-Lobato,
Andreas Bender,
Sergio Bacallado
Abstract:
The field of machine learning for drug discovery is witnessing an explosion of novel methods. These methods are often benchmarked on simple physicochemical properties such as solubility or general druglikeness, which can be readily computed. However, these properties are poor representatives of objective functions in drug design, mainly because they do not depend on the candidate's interaction wit…
▽ More
The field of machine learning for drug discovery is witnessing an explosion of novel methods. These methods are often benchmarked on simple physicochemical properties such as solubility or general druglikeness, which can be readily computed. However, these properties are poor representatives of objective functions in drug design, mainly because they do not depend on the candidate's interaction with the target. By contrast, molecular docking is a widely successful method in drug discovery to estimate binding affinities. However, docking simulations require a significant amount of domain knowledge to set up correctly which hampers adoption. To this end, we present DOCKSTRING, a bundle for meaningful and robust comparison of ML models consisting of three components: (1) an open-source Python package for straightforward computation of docking scores; (2) an extensive dataset of docking scores and poses of more than 260K ligands for 58 medically-relevant targets; and (3) a set of pharmaceutically-relevant benchmark tasks including regression, virtual screening, and de novo design. The Python package implements a robust ligand and target preparation protocol that allows non-experts to obtain meaningful docking scores. Our dataset is the first to include docking poses, as well as the first of its size that is a full matrix, thus facilitating experiments in multiobjective optimization and transfer learning. Overall, our results indicate that docking scores are a more appropriate evaluation objective than simple physicochemical properties, yielding more realistic benchmark tasks and molecular candidates.
△ Less
Submitted 28 October, 2021;
originally announced October 2021.
-
Characteristic polynomials of random truncations: moments, duality and asymptotics
Authors:
Alexander Serebryakov,
Nick Simm,
Guillaume Dubach
Abstract:
We study moments of characteristic polynomials of truncated Haar distributed matrices from the three classical compact groups O(N), U(N) and Sp(2N). For finite matrix size we calculate the moments in terms of hypergeometric functions of matrix argument and give explicit integral representations highlighting the duality between the moment and the matrix size as well as the duality between the ortho…
▽ More
We study moments of characteristic polynomials of truncated Haar distributed matrices from the three classical compact groups O(N), U(N) and Sp(2N). For finite matrix size we calculate the moments in terms of hypergeometric functions of matrix argument and give explicit integral representations highlighting the duality between the moment and the matrix size as well as the duality between the orthogonal and symplectic cases. Asymptotic expansions in strong and weak non-unitarity regimes are obtained. Using the connection to matrix hypergeometric functions, we establish limit theorems for the log-modulus of the characteristic polynomial evaluated on the unit circle.
△ Less
Submitted 15 November, 2021; v1 submitted 21 September, 2021;
originally announced September 2021.
-
Fluctuations and correlations for products of real asymmetric random matrices
Authors:
Will FitzGerald,
Nick Simm
Abstract:
We study the real eigenvalue statistics of products of independent real Ginibre random matrices. These are matrices all of whose entries are real i.i.d. standard Gaussian random variables. For such product ensembles, we demonstrate the asymptotic normality of suitably normalised linear statistics of the real eigenvalues and compute the limiting variance explicitly in both global and mesoscopic reg…
▽ More
We study the real eigenvalue statistics of products of independent real Ginibre random matrices. These are matrices all of whose entries are real i.i.d. standard Gaussian random variables. For such product ensembles, we demonstrate the asymptotic normality of suitably normalised linear statistics of the real eigenvalues and compute the limiting variance explicitly in both global and mesoscopic regimes. A key part of our proof establishes uniform decorrelation estimates for the related Pfaffian point process, thereby allowing us to exploit weak dependence of the real eigenvalues to give simple and quick proofs of the central limit theorems under quite general conditions. We also establish the universality of these point processes. We compute the asymptotic limit of all correlation functions of the real eigenvalues in the bulk, origin and spectral edge regimes. By a suitable strengthening of the convergence at the edge, we also obtain the limiting fluctuations of the largest real eigenvalue. Near the origin we find new limiting distributions characterising the smallest positive real eigenvalue.
△ Less
Submitted 1 September, 2021;
originally announced September 2021.
-
On the number of real eigenvalues of a product of truncated orthogonal random matrices
Authors:
Alex Little,
Francesco Mezzadri,
Nick Simm
Abstract:
Let $O$ be chosen uniformly at random from the group of $(N+L) \times (N+L)$ orthogonal matrices. Denote by $\tilde{O}$ the upper-left $N \times N$ corner of $O$, which we refer to as a truncation of $O$. In this paper we prove two conjectures of Forrester, Ipsen and Kumar (2020) on the number of real eigenvalues $N^{(m)}_{\mathbb{R}}$ of the product matrix $\tilde{O}_{1}\ldots \tilde{O}_{m}$, whe…
▽ More
Let $O$ be chosen uniformly at random from the group of $(N+L) \times (N+L)$ orthogonal matrices. Denote by $\tilde{O}$ the upper-left $N \times N$ corner of $O$, which we refer to as a truncation of $O$. In this paper we prove two conjectures of Forrester, Ipsen and Kumar (2020) on the number of real eigenvalues $N^{(m)}_{\mathbb{R}}$ of the product matrix $\tilde{O}_{1}\ldots \tilde{O}_{m}$, where the matrices $\{\tilde{O}_{j}\}_{j=1}^{m}$ are independent copies of $\tilde{O}$. When $L$ grows in proportion to $N$, we prove that $$ \mathbb{E}(N^{(m)}_{\mathbb{R}}) = \sqrt{\frac{2m L}π}\,\mathrm{arctanh}\left(\sqrt{\frac{N}{N+L}}\right) + O(1), \qquad N \to \infty. $$ We also prove the conjectured form of the limiting real eigenvalue distribution of the product matrix. Finally, we consider the opposite regime where $L$ is fixed with respect to $N$, known as the regime of weak non-orthogonality. In this case each matrix in the product is very close to an orthogonal matrix. We show that $\mathbb{E}(N^{(m)}_{\mathbb{R}}) \sim c_{L,m}\,\log(N)$ as $N \to \infty$ and compute the constant $c_{L,m}$ explicitly. These results generalise the known results in the one matrix case due to Khoruzhenko, Sommers and Życzkowski (2010).
△ Less
Submitted 17 February, 2021;
originally announced February 2021.
-
Symmetry-Aware Actor-Critic for 3D Molecular Design
Authors:
Gregor N. C. Simm,
Robert Pinsler,
Gábor Csányi,
José Miguel Hernández-Lobato
Abstract:
Automating molecular design using deep reinforcement learning (RL) has the potential to greatly accelerate the search for novel materials. Despite recent progress on leveraging graph representations to design molecules, such methods are fundamentally limited by the lack of three-dimensional (3D) information. In light of this, we propose a novel actor-critic architecture for 3D molecular design tha…
▽ More
Automating molecular design using deep reinforcement learning (RL) has the potential to greatly accelerate the search for novel materials. Despite recent progress on leveraging graph representations to design molecules, such methods are fundamentally limited by the lack of three-dimensional (3D) information. In light of this, we propose a novel actor-critic architecture for 3D molecular design that can generate molecular structures unattainable with previous approaches. This is achieved by exploiting the symmetries of the design process through a rotationally covariant state-action representation based on a spherical harmonics series expansion. We demonstrate the benefits of our approach on several 3D molecular design tasks, where we find that building in such symmetries significantly improves generalization and the quality of generated molecules.
△ Less
Submitted 25 November, 2020;
originally announced November 2020.
-
Secular Coefficients and the Holomorphic Multiplicative Chaos
Authors:
Joseph Najnudel,
Elliot Paquette,
Nick Simm
Abstract:
We study the secular coefficients of $N \times N$ random unitary matrices $U_{N}$ drawn from the Circular $β$-Ensemble, which are defined as the coefficients of $\{z^n\}$ in the characteristic polynomial $\det(1-zU_{N}^{*})$. When $β> 4$ we obtain a new class of limiting distributions that arise when both $n$ and $N$ tend to infinity simultaneously. We solve an open problem of Diaconis and Gamburd…
▽ More
We study the secular coefficients of $N \times N$ random unitary matrices $U_{N}$ drawn from the Circular $β$-Ensemble, which are defined as the coefficients of $\{z^n\}$ in the characteristic polynomial $\det(1-zU_{N}^{*})$. When $β> 4$ we obtain a new class of limiting distributions that arise when both $n$ and $N$ tend to infinity simultaneously. We solve an open problem of Diaconis and Gamburd by showing that for $β=2$, the middle coefficient tends to zero as $N \to \infty$. We show how the theory of Gaussian multiplicative chaos (GMC) plays a prominent role in these problems and in the explicit description of the obtained limiting distributions. We extend the remarkable magic square formula of Diaconis and Gamburd for the moments of secular coefficients to all $β>0$ and analyse the asymptotic behaviour of the moments. We obtain estimates on the order of magnitude of the secular coefficients for all $β> 0,$ and these estimates are sharp when $β\geq 2$. These insights motivated us to introduce a new stochastic object associated with the secular coefficients, which we call Holomorphic Multiplicative Chaos (HMC). Viewing the HMC as a random distribution, we prove a sharp result about its regularity in an appropriate Sobolev space. Our proofs expose and exploit several novel connections with other areas, including random permutations, Tauberian theorems and combinatorics.
△ Less
Submitted 3 November, 2020;
originally announced November 2020.
-
Reinforcement Learning for Molecular Design Guided by Quantum Mechanics
Authors:
Gregor N. C. Simm,
Robert Pinsler,
José Miguel Hernández-Lobato
Abstract:
Automating molecular design using deep reinforcement learning (RL) holds the promise of accelerating the discovery of new chemical compounds. Existing approaches work with molecular graphs and thus ignore the location of atoms in space, which restricts them to 1) generating single organic molecules and 2) heuristic reward functions. To address this, we present a novel RL formulation for molecular…
▽ More
Automating molecular design using deep reinforcement learning (RL) holds the promise of accelerating the discovery of new chemical compounds. Existing approaches work with molecular graphs and thus ignore the location of atoms in space, which restricts them to 1) generating single organic molecules and 2) heuristic reward functions. To address this, we present a novel RL formulation for molecular design in Cartesian coordinates, thereby extending the class of molecules that can be built. Our reward function is directly based on fundamental physical properties such as the energy, which we approximate via fast quantum-chemical methods. To enable progress towards de-novo molecular design, we introduce MolGym, an RL environment comprising several challenging molecular design tasks along with baselines. In our experiments, we show that our agent can efficiently learn to solve these tasks from scratch by working in a translation and rotation invariant state-action space.
△ Less
Submitted 29 June, 2020; v1 submitted 18 February, 2020;
originally announced February 2020.
-
A Generative Model for Molecular Distance Geometry
Authors:
Gregor N. C. Simm,
José Miguel Hernández-Lobato
Abstract:
Great computational effort is invested in generating equilibrium states for molecular systems using, for example, Markov chain Monte Carlo. We present a probabilistic model that generates statistically independent samples for molecules from their graph representations. Our model learns a low-dimensional manifold that preserves the geometry of local atomic neighborhoods through a principled learnin…
▽ More
Great computational effort is invested in generating equilibrium states for molecular systems using, for example, Markov chain Monte Carlo. We present a probabilistic model that generates statistically independent samples for molecules from their graph representations. Our model learns a low-dimensional manifold that preserves the geometry of local atomic neighborhoods through a principled learning representation that is based on Euclidean distance geometry. In a new benchmark for molecular conformation generation, we show experimentally that our generative model achieves state-of-the-art accuracy. Finally, we show how to use our model as a proposal distribution in an importance sampling scheme to compute molecular properties.
△ Less
Submitted 13 August, 2020; v1 submitted 25 September, 2019;
originally announced September 2019.
-
Systematic Microsolvation Approach with a Cluster-Continuum Scheme and Conformational Sampling
Authors:
Gregor N. Simm,
Paul L. Türtscher,
Markus Reiher
Abstract:
Solvation is a notoriously difficult and nagging problem for the rigorous theoretical description of chemistry in the liquid phase. Successes and failures of various approaches ranging from implicit solvation modeling through dielectric continuum embedding and microsolvated quantum chemical modeling to explicit molecular dynamics highlight this situation. Here, we focus on quantum chemical microso…
▽ More
Solvation is a notoriously difficult and nagging problem for the rigorous theoretical description of chemistry in the liquid phase. Successes and failures of various approaches ranging from implicit solvation modeling through dielectric continuum embedding and microsolvated quantum chemical modeling to explicit molecular dynamics highlight this situation. Here, we focus on quantum chemical microsolvation and discuss an explicit conformational sampling ansatz to make this approach systematic. For this purpose, we introduce an algorithm for the rolling and automated microsolvation of solutes. Our protocol takes conformational sampling and rearrangements in the solvent shell into account. Its reliability is assessed by monitoring the evolution of the spread and average of the observables of interest.
△ Less
Submitted 27 January, 2020; v1 submitted 14 September, 2019;
originally announced September 2019.
-
Characteristic polynomials of complex random matrices and Painlevé transcendents
Authors:
Alfredo Deaño,
Nick Simm
Abstract:
We study expectations of powers and correlation functions for characteristic polynomials of $N \times N$ non-Hermitian random matrices. For the $1$-point and $2$-point correlation function, we obtain several characterizations in terms of Painlevé transcendents, both at finite-$N$ and asymptotically as $N \to \infty$. In the asymptotic analysis, two regimes of interest are distinguished: boundary a…
▽ More
We study expectations of powers and correlation functions for characteristic polynomials of $N \times N$ non-Hermitian random matrices. For the $1$-point and $2$-point correlation function, we obtain several characterizations in terms of Painlevé transcendents, both at finite-$N$ and asymptotically as $N \to \infty$. In the asymptotic analysis, two regimes of interest are distinguished: boundary asymptotics where parameters of the correlation function can touch the boundary of the limiting eigenvalue support and bulk asymptotics where they are strictly inside the support. For the complex Ginibre ensemble this involves Painlevé IV at the boundary as $N \to \infty$. Our approach, together with the results in \cite{HW17} suggests that this should arise in a much broader class of planar models. For the bulk asymptotics, one of our results can be interpreted as the merging of two `planar Fisher-Hartwig singularities' where Painlevé V arises in the asymptotics. We also discuss the correspondence of our results with a normal matrix model with $d$-fold rotational symmetries known as the \textit{lemniscate ensemble}, recently studied in \cite{BGM, BGG18}. Our approach is flexible enough to apply to non-Gaussian models such as the truncated unitary ensemble or induced Ginibre ensemble; we show that in the former case Painlevé VI arises at finite-$N$. Scaling near the boundary leads to Painlevé V, in contrast to the Ginibre ensemble.
△ Less
Submitted 1 June, 2020; v1 submitted 13 September, 2019;
originally announced September 2019.
-
Exploration of Reaction Pathways and Chemical Transformation Networks
Authors:
Gregor N. Simm,
Alain C. Vaucher,
Markus Reiher
Abstract:
For the investigation of chemical reaction networks, the identification of all relevant intermediates and elementary reactions is mandatory. Many algorithmic approaches exist that perform explorations efficiently and automatedly. These approaches differ in their application range, the level of completeness of the exploration, as well as the amount of heuristics and human intervention required. Her…
▽ More
For the investigation of chemical reaction networks, the identification of all relevant intermediates and elementary reactions is mandatory. Many algorithmic approaches exist that perform explorations efficiently and automatedly. These approaches differ in their application range, the level of completeness of the exploration, as well as the amount of heuristics and human intervention required. Here, we describe and compare the different approaches based on these criteria. Future directions leveraging the strengths of chemical heuristics, human interaction, and physical rigor are discussed.
△ Less
Submitted 3 December, 2018; v1 submitted 17 October, 2018;
originally announced October 2018.
-
Error-Controlled Exploration of Chemical Reaction Networks with Gaussian Processes
Authors:
Gregor N. Simm,
Markus Reiher
Abstract:
For a theoretical understanding of the reactivity of complex chemical systems, relative energies of stationary points on potential energy hypersurfaces need to be calculated to high accuracy. Due to the large number of intermediates present in all but the simplest chemical processes, approximate quantum chemical methods are required that allow for fast evaluations of the relative energies, but at…
▽ More
For a theoretical understanding of the reactivity of complex chemical systems, relative energies of stationary points on potential energy hypersurfaces need to be calculated to high accuracy. Due to the large number of intermediates present in all but the simplest chemical processes, approximate quantum chemical methods are required that allow for fast evaluations of the relative energies, but at the expense of accuracy. Despite the plethora of benchmark studies, the accuracy of a quantum chemical method is often difficult to assess. Moreover, a significant improvement of a method's accuracy (e.g., through reparameterization or systematic model extension) is rarely possible. Here, we present a new approach that allows for the systematic, problem-oriented, and rolling improvement of quantum chemical results through the application of Gaussian processes. Due to its Bayesian nature, reliable error estimates are provided for each prediction. A reference method of high accuracy can be employed, if the uncertainty associated with a particular calculation is above a given threshold. The new data point is then added to a growing data set in order to continuously improve the model, and as a result, all subsequent predictions. Previous predictions are validated by the updated model to ensure that uncertainties remain within the given confidence bound, which we call backtracking. We demonstrate our approach at the example of a complex chemical reaction network.
△ Less
Submitted 27 October, 2018; v1 submitted 24 May, 2018;
originally announced May 2018.
-
Moments of random matrices and hypergeometric orthogonal polynomials
Authors:
Fabio Deelan Cunden,
Francesco Mezzadri,
Neil O'Connell,
Nick Simm
Abstract:
We establish a new connection between moments of $n \times n$ random matrices $X_n$ and hypergeometric orthogonal polynomials. Specifically, we consider moments $\mathbb{E}\mathrm{Tr} X_n^{-s}$ as a function of the complex variable $s \in \mathbb{C}$, whose analytic structure we describe completely. We discover several remarkable features, including a reflection symmetry (or functional equation),…
▽ More
We establish a new connection between moments of $n \times n$ random matrices $X_n$ and hypergeometric orthogonal polynomials. Specifically, we consider moments $\mathbb{E}\mathrm{Tr} X_n^{-s}$ as a function of the complex variable $s \in \mathbb{C}$, whose analytic structure we describe completely. We discover several remarkable features, including a reflection symmetry (or functional equation), zeros on a critical line in the complex plane, and orthogonality relations. An application of the theory resolves part of an integrality conjecture of Cunden et al. [F. D. Cunden, F. Mezzadri, N. J. Simm and P. Vivo, J. Math. Phys. 57 (2016)] on the time-delay matrix of chaotic cavities. In each of the classical ensembles of random matrix theory (Gaussian, Laguerre, Jacobi) we characterise the moments in terms of the Askey scheme of hypergeometric orthogonal polynomials. We also calculate the leading order $n\to\infty$ asymptotics of the moments and discuss their symmetries and zeroes. We discuss aspects of these phenomena beyond the random matrix setting, including the Mellin transform of products and Wronskians of pairs of classical orthogonal polynomials. When the random matrix model has orthogonal or symplectic symmetry, we obtain a new duality formula relating their moments to hypergeometric orthogonal polynomials.
△ Less
Submitted 11 January, 2019; v1 submitted 22 May, 2018;
originally announced May 2018.
-
Context-Driven Exploration of Complex Chemical Reaction Networks
Authors:
Gregor N. Simm,
Markus Reiher
Abstract:
The construction of a reaction network containing all relevant intermediates and elementary reactions is necessary for the accurate description of chemical processes. In the case of a complex chemical reaction (involving, for instance, many reactants or highly reactive species), the size of such network may grow rapidly. Here, we present a computational protocol that constructs such reaction netwo…
▽ More
The construction of a reaction network containing all relevant intermediates and elementary reactions is necessary for the accurate description of chemical processes. In the case of a complex chemical reaction (involving, for instance, many reactants or highly reactive species), the size of such network may grow rapidly. Here, we present a computational protocol that constructs such reaction networks in a fully automated fashion steered in an intuitive, graph-based fashion through a single graphical user interface. Starting from a set of initial reagents new intermediates are explored through intra- and intermolecular reactions of already explored intermediates or new reactants presented to the network. This is done by assembling reactive complexes based on heuristic rules derived from conceptual electronic-structure theory and exploring the corresponding approximate reaction path. A subsequent path refinement leads to a minimum-energy path which connects the new intermediate to the existing ones to form a connected reaction network. Tree traversal algorithms are then employed to detect reaction channels and catalytic cycles. We apply our protocol to the formose reaction to study different pathways of sugar formation and to rationalize its autocatalytic nature.
△ Less
Submitted 25 October, 2017; v1 submitted 7 September, 2017;
originally announced September 2017.
-
Error Assessment of Computational Models in Chemistry
Authors:
Gregor N. Simm,
Jonny Proppe,
Markus Reiher
Abstract:
Computational models in chemistry rely on a number of approximations. The effect of such approximations on observables derived from them is often unpredictable. Therefore, it is challenging to quantify the uncertainty of a computational result, which, however, is necessary to assess the suitability of a computational model. Common performance statistics such as the mean absolute error are prone to…
▽ More
Computational models in chemistry rely on a number of approximations. The effect of such approximations on observables derived from them is often unpredictable. Therefore, it is challenging to quantify the uncertainty of a computational result, which, however, is necessary to assess the suitability of a computational model. Common performance statistics such as the mean absolute error are prone to failure as they do not distinguish the explainable (systematic) part of the errors from their unexplainable (random) part. In this paper, we discuss problems and solutions for performance assessment of computational models based on several examples from the quantum chemistry literature. For this purpose, we elucidate the different sources of uncertainty, the elimination of systematic errors, and the combination of individual uncertainty components to the uncertainty of a prediction.
△ Less
Submitted 2 February, 2017;
originally announced February 2017.
-
On the real spectrum of a product of Gaussian random matrices
Authors:
Nick Simm
Abstract:
Let $X_{m} = G_{1}\ldots G_{m}$ denote the product of $m$ independent random matrices of size $N \times N$, with each matrix in the product consisting of independent standard Gaussian variables. Denoting by $N_{\mathbb{R}}(m)$ the total number of real eigenvalues of $X_{m}$, we show that for $m$ fixed \begin{equation*} \mathbb{E}(N_{\mathbb{R}}(m)) = \sqrt{\frac{2Nm}π}+O(\log(N)), \qquad N \to \in…
▽ More
Let $X_{m} = G_{1}\ldots G_{m}$ denote the product of $m$ independent random matrices of size $N \times N$, with each matrix in the product consisting of independent standard Gaussian variables. Denoting by $N_{\mathbb{R}}(m)$ the total number of real eigenvalues of $X_{m}$, we show that for $m$ fixed \begin{equation*} \mathbb{E}(N_{\mathbb{R}}(m)) = \sqrt{\frac{2Nm}π}+O(\log(N)), \qquad N \to \infty. \end{equation*} This generalizes a well-known result of Edelman et al. \cite{EKS94} to all $m>1$. Furthermore, we show that the normalized global density of real eigenvalues converges weakly in expectation to the density of the random variable $|U|^{m}B$ where $U$ is uniform on $[-1,1]$ and $B$ is Bernoulli on $\{-1,1\}$. This proves a conjecture of Forrester and Ipsen \cite{FI16}. The results are obtained by the asymptotic analysis of a certain Meijer G-function.
△ Less
Submitted 31 January, 2017;
originally announced January 2017.
-
Subcritical multiplicative chaos for regularized counting statistics from random matrix theory
Authors:
Gaultier Lambert,
Dmitry Ostrovsky,
Nick Simm
Abstract:
For an $N \times N$ random unitary matrix $U_N$, we consider the random field defined by counting the number of eigenvalues of $U_N$ in a mesoscopic arc of the unit circle, regularized at an $N$-dependent scale $ε_N>0$. We prove that the renormalized exponential of this field converges as $N \to \infty$ to a Gaussian multiplicative chaos measure in the whole subcritical phase. In addition, we show…
▽ More
For an $N \times N$ random unitary matrix $U_N$, we consider the random field defined by counting the number of eigenvalues of $U_N$ in a mesoscopic arc of the unit circle, regularized at an $N$-dependent scale $ε_N>0$. We prove that the renormalized exponential of this field converges as $N \to \infty$ to a Gaussian multiplicative chaos measure in the whole subcritical phase. In addition, we show that the moments of the total mass converge to a Selberg-like integral and by taking a further limit as the size of the arc diverges, we establish part of the conjectures in \cite{Ost16}. By an analogous construction, we prove that the multiplicative chaos measure coming from the sine process has the same distribution, which strongly suggests that this limiting object should be universal. The proofs are based on the asymptotic analysis of certain Toeplitz or Fredholm determinants using the Borodin-Okounkov formula or a Riemann-Hilbert problem for integrable operators. Our approach to the $L^{1}$-phase is based on a generalization of the construction in Berestycki \cite{Berestycki15} to random fields which are only \textit{asymptotically} Gaussian. In particular, our method could have applications to other random fields coming from either random matrix theory or a different context.
△ Less
Submitted 11 April, 2018; v1 submitted 7 December, 2016;
originally announced December 2016.
-
On the probability of positive-definiteness in the gGUE via semi-classical Laguerre polynomials
Authors:
Alfredo Deaño,
Nick Simm
Abstract:
In this paper, we compute the probability that an $N \times N$ matrix from the generalised Gaussian Unitary Ensemble (gGUE) is positive definite, extending a previous result of Dean and Majumdar \cite{DM}. For this purpose, we work out the large degree asymptotics of semi-classical Laguerre polynomials and their recurrence coefficients, using the steepest descent analysis of the corresponding Riem…
▽ More
In this paper, we compute the probability that an $N \times N$ matrix from the generalised Gaussian Unitary Ensemble (gGUE) is positive definite, extending a previous result of Dean and Majumdar \cite{DM}. For this purpose, we work out the large degree asymptotics of semi-classical Laguerre polynomials and their recurrence coefficients, using the steepest descent analysis of the corresponding Riemann--Hilbert problem.
△ Less
Submitted 23 May, 2017; v1 submitted 26 October, 2016;
originally announced October 2016.
-
Large-$N$ expansion for the time-delay matrix of ballistic chaotic cavities
Authors:
Fabio Deelan Cunden,
Francesco Mezzadri,
Nick Simm,
Pierpaolo Vivo
Abstract:
We consider the $1/N$-expansion of the moments of the proper delay times for a ballistic chaotic cavity supporting $N$ scattering channels. In the random matrix approach, these moments correspond to traces of negative powers of Wishart matrices. For systems with and without broken time reversal symmetry (Dyson indices $β=1$ and $β=2$) we obtain a recursion relation, which efficiently generates the…
▽ More
We consider the $1/N$-expansion of the moments of the proper delay times for a ballistic chaotic cavity supporting $N$ scattering channels. In the random matrix approach, these moments correspond to traces of negative powers of Wishart matrices. For systems with and without broken time reversal symmetry (Dyson indices $β=1$ and $β=2$) we obtain a recursion relation, which efficiently generates the coefficients of the $1/N$-expansion of the moments. The integrality of these coefficients and their possible diagrammatic interpretation is discussed.
△ Less
Submitted 27 October, 2016; v1 submitted 1 July, 2016;
originally announced July 2016.
-
Systematic Error Estimation for Chemical Reaction Energies
Authors:
Gregor N. Simm,
Markus Reiher
Abstract:
For the theoretical understanding of the reactivity of complex chemical systems accurate relative energies between intermediates and transition states are required. Despite its popularity, density functional theory (DFT) often fails to provide sufficiently accurate data, especially for molecules containing transition metals. Due to the huge number of intermediates that need to be studied for all b…
▽ More
For the theoretical understanding of the reactivity of complex chemical systems accurate relative energies between intermediates and transition states are required. Despite its popularity, density functional theory (DFT) often fails to provide sufficiently accurate data, especially for molecules containing transition metals. Due to the huge number of intermediates that need to be studied for all but the simplest chemical processes, DFT is to date the only method that is computationally feasible. Here, we present a Bayesian framework for DFT that allows for error estimation of calculated properties. Since the optimal choice of parameters in present-day density functionals is strongly system dependent, we advocate for a system-focused re-parameterization. While, at first sight, this approach conflicts with the first-principles character of DFT that should make it in principle system independent, we deliberately introduce system dependence because we can then assign a stochastically meaningful error to the system-dependent parametrization that makes it non-arbitrary. By re-parameterizing a functional that was derived on a sound physical basis to a chemical system of interest we obtain a functional that yields reliable confidence intervals for reaction energies. We demonstrate our approach at the example of catalytic nitrogen fixation.
△ Less
Submitted 28 March, 2016;
originally announced March 2016.
-
Correlators for the Wigner-Smith time-delay matrix of chaotic cavities
Authors:
Fabio Deelan Cunden,
Francesco Mezzadri,
Nick Simm,
Pierpaolo Vivo
Abstract:
We study the Wigner-Smith time-delay matrix $Q$ of a ballistic quantum dot supporting $N$ scattering channels. We compute the $v$-point correlators of the power traces $\mathrm{Tr} Q^κ$ for arbitrary $v\geq1$ at leading order for large $N$ using techniques from the random matrix theory approach to quantum chromodynamics. We conjecture that the cumulants of the $\mathrm{Tr} Q^κ$'s are integer-value…
▽ More
We study the Wigner-Smith time-delay matrix $Q$ of a ballistic quantum dot supporting $N$ scattering channels. We compute the $v$-point correlators of the power traces $\mathrm{Tr} Q^κ$ for arbitrary $v\geq1$ at leading order for large $N$ using techniques from the random matrix theory approach to quantum chromodynamics. We conjecture that the cumulants of the $\mathrm{Tr} Q^κ$'s are integer-valued at leading order in $N$ and include a MATHEMATICA code that computes their generating functions recursively.
△ Less
Submitted 20 March, 2016; v1 submitted 25 January, 2016;
originally announced January 2016.
-
Central limit theorems for the real eigenvalues of large Gaussian random matrices
Authors:
N. J. Simm
Abstract:
Let $G$ be an $N \times N$ real matrix whose entries are independent identically distributed standard normal random variables $G_{ij} \sim \mathcal{N}(0,1)$. The eigenvalues of such matrices are known to form a two-component system consisting of purely real and complex conjugated points. The purpose of this note is to show that by appropriately adapting the methods of \cite{KPTTZ15}, we can prove…
▽ More
Let $G$ be an $N \times N$ real matrix whose entries are independent identically distributed standard normal random variables $G_{ij} \sim \mathcal{N}(0,1)$. The eigenvalues of such matrices are known to form a two-component system consisting of purely real and complex conjugated points. The purpose of this note is to show that by appropriately adapting the methods of \cite{KPTTZ15}, we can prove a central limit theorem of the following form: if $λ_{1},\ldots,λ_{N_{\mathbb{R}}}$ are the real eigenvalues of $G$, then for any even polynomial function $P(x)$ and even $N=2n$, we have the convergence in distribution to a normal random variable
\begin{equation}
\frac{1}{\sqrt{\mathbb{E}(N_{\mathbb{R}})}}\left(\sum_{j=1}^{N_{\mathbb{R}}}P(λ_{j})-\mathbb{E}\sum_{j=1}^{N_{\mathbb{R}}}P(λ_{j})\right) \to \mathcal{N}(0,σ^{2}(P))
\end{equation} as $n \to \infty$, where $σ^{2}(P) = \frac{2-\sqrt{2}}{2}\int_{-1}^{1}P(x)^{2}\,dx$.
△ Less
Submitted 4 December, 2015;
originally announced December 2015.
-
A Bayesian Approach to Calibrating High-Throughput Virtual Screening Results and Application to Organic Photovoltaic Materials
Authors:
Edward O. Pyzer-Knapp,
Gregor N. Simm,
Alan Aspuru-Guzik
Abstract:
A novel approach for calibrating quantum-chemical properties determined as part of a high-throughput virtual screen to experimental analogs is presented. Information on the molecular graph is extracted through the use of extended connectivity fingerprints, and exploited using a Gaussian process to calibrate both electronic properties such as frontier orbital energies, and optical gaps and device p…
▽ More
A novel approach for calibrating quantum-chemical properties determined as part of a high-throughput virtual screen to experimental analogs is presented. Information on the molecular graph is extracted through the use of extended connectivity fingerprints, and exploited using a Gaussian process to calibrate both electronic properties such as frontier orbital energies, and optical gaps and device properties such as short circuit current density, open circuit voltage and power conversion efficiency. The Bayesian nature of this process affords a value for uncertainty in addition to each calibrated value. This allows the researcher to gain intuition about the model as well as the ability to respect its bounds.
△ Less
Submitted 1 October, 2015;
originally announced October 2015.
-
Heuristics-Guided Exploration of Reaction Mechanisms
Authors:
Maike Bergeler,
Gregor N. Simm,
Jonny Proppe,
Markus Reiher
Abstract:
For the investigation of chemical reaction networks, the efficient and accurate determination of all relevant intermediates and elementary reactions is mandatory. The complexity of such a network may grow rapidly, in particular if reactive species are involved that might cause a myriad of side reactions. Without automation, a complete investigation of complex reaction mechanisms is tedious and pos…
▽ More
For the investigation of chemical reaction networks, the efficient and accurate determination of all relevant intermediates and elementary reactions is mandatory. The complexity of such a network may grow rapidly, in particular if reactive species are involved that might cause a myriad of side reactions. Without automation, a complete investigation of complex reaction mechanisms is tedious and possibly unfeasible. Therefore, only the expected dominant reaction paths of a chemical reaction network (e.g., a catalytic cycle or an enzymatic cascade) are usually explored in practice. Here, we present a computational protocol that constructs such networks in a parallelized and automated manner. Molecular structures of reactive complexes are generated based on heuristic rules derived from conceptual electronic-structure theory and subsequently optimized by quantum chemical methods to produce stable intermediates of an emerging reaction network. Pairs of intermediates in this network that might be related by an elementary reaction according to some structural similarity measure are then automatically detected and subjected to an automated search for the connecting transition state. The results are visualized as an automatically generated network graph, from which a comprehensive picture of the mechanism of a complex chemical process can be obtained that greatly facilitates the analysis of the whole network. We apply our protocol to the Schrock dinitrogen-fixation catalyst to study alternative pathways of catalytic ammonia production.
△ Less
Submitted 23 October, 2015; v1 submitted 10 September, 2015;
originally announced September 2015.
-
On the distribution of maximum value of the characteristic polynomial of GUE random matrices
Authors:
Yan V. Fyodorov,
Nicholas J. Simm
Abstract:
Motivated by recently discovered relations between logarithmically correlated Gaussian processes and characteristic polynomials of large random $N \times N$ matrices $H$ from the Gaussian Unitary Ensemble (GUE), we consider the problem of characterising the distribution of the global maximum of $D_{N}(x):=-\log|\det(xI-H)|$ as $N \to \infty$ and $x\in (-1,1)$. We arrive at an explicit expression f…
▽ More
Motivated by recently discovered relations between logarithmically correlated Gaussian processes and characteristic polynomials of large random $N \times N$ matrices $H$ from the Gaussian Unitary Ensemble (GUE), we consider the problem of characterising the distribution of the global maximum of $D_{N}(x):=-\log|\det(xI-H)|$ as $N \to \infty$ and $x\in (-1,1)$. We arrive at an explicit expression for the asymptotic probability density of the (appropriately shifted) maximum by combining the rigorous Fisher-Hartwig asymptotics due to Krasovsky \cite{K07} with the heuristic {\it freezing transition} scenario for logarithmically correlated processes. Although the general idea behind the method is the same as for the earlier considered case of the Circular Unitary Ensemble, the present GUE case poses new challenges. In particular we show how the conjectured {\it self-duality} in the freezing scenario plays the crucial role in our selection of the form of the maximum distribution. Finally, we demonstrate a good agreement of the found probability density with the results of direct numerical simulations of the maxima of $D_{N}(x)$.
△ Less
Submitted 21 June, 2015; v1 submitted 24 March, 2015;
originally announced March 2015.
-
Mesoscopic linear statistics of Wigner matrices
Authors:
A. Lodhia,
N. J. Simm
Abstract:
We study linear spectral statistics of $N \times N$ Wigner random matrices $\mathcal{H}$ on mesoscopic scales. Under mild assumptions on the matrix entries of $\mathcal{H}$, we prove that after centering and normalizing, the trace of the resolvent $\mathrm{Tr}(\mathcal{H}-z)^{-1}$ converges to a stationary Gaussian process as $N \to \infty$ on scales $N^{-1/3} \ll \mathrm{Im}(z) \ll 1$ and explici…
▽ More
We study linear spectral statistics of $N \times N$ Wigner random matrices $\mathcal{H}$ on mesoscopic scales. Under mild assumptions on the matrix entries of $\mathcal{H}$, we prove that after centering and normalizing, the trace of the resolvent $\mathrm{Tr}(\mathcal{H}-z)^{-1}$ converges to a stationary Gaussian process as $N \to \infty$ on scales $N^{-1/3} \ll \mathrm{Im}(z) \ll 1$ and explicitly compute the covariance structure. The limit process is related to certain regularizations of fractional Brownian motion and logarithmically correlated fields appearing in \cite{FKS13}. Finally, we extend our results to general mesoscopic linear statistics and prove that the limiting covariance is given by the $H^{1/2}$-norm of the test functions.
△ Less
Submitted 11 March, 2015;
originally announced March 2015.
-
Energy Landscape of the Finite-Size Mean-field 2-Spin Spherical Model and Topology Trivialization
Authors:
Dhagash Mehta,
Jonathan D. Hauenstein,
Matthew Niemerg,
Nicholas J. Simm,
Daniel A. Stariolo
Abstract:
Motivated by the recently observed phenomenon of topology trivialization of potential energy landscapes (PELs) for several statistical mechanics models, we perform a numerical study of the finite size $2$-spin spherical model using both numerical polynomial homotopy continuation and a reformulation via non-hermitian matrices. The continuation approach computes all of the complex stationary points…
▽ More
Motivated by the recently observed phenomenon of topology trivialization of potential energy landscapes (PELs) for several statistical mechanics models, we perform a numerical study of the finite size $2$-spin spherical model using both numerical polynomial homotopy continuation and a reformulation via non-hermitian matrices. The continuation approach computes all of the complex stationary points of this model while the matrix approach computes the real stationary points. Using these methods, we compute the average number of stationary points while changing the topology of the PEL as well as the variance. Histograms of these stationary points are presented along with an analysis regarding the complex stationary points. This work connects topology trivialization to two different branches of mathematics: algebraic geometry and catastrophe theory, which is fertile ground for further interdisciplinary research.
△ Less
Submitted 29 September, 2014;
originally announced September 2014.
-
Fractional Brownian motion with Hurst index $H=0$ and the Gaussian Unitary Ensemble
Authors:
Y. V. Fyodorov,
B. A. Khoruzhenko,
N. J. Simm
Abstract:
The goal of this paper is to establish a relation between characteristic polynomials of $N\times N$ GUE random matrices $\mathcal{H}$ as $N\to\infty$, and Gaussian processes with logarithmic correlations. We introduce a regularized version of fractional Brownian motion with zero Hurst index, which is a Gaussian process with stationary increments and logarithmic increment structure. Then we prove t…
▽ More
The goal of this paper is to establish a relation between characteristic polynomials of $N\times N$ GUE random matrices $\mathcal{H}$ as $N\to\infty$, and Gaussian processes with logarithmic correlations. We introduce a regularized version of fractional Brownian motion with zero Hurst index, which is a Gaussian process with stationary increments and logarithmic increment structure. Then we prove that this process appears as a limit of $D_N(z)=-\log|\det(\mathcal{H}-zI)|$ on mesoscopic scales as $N\to\infty$. By employing a Fourier integral representation, we use this to prove a continuous analogue of a result by Diaconis and Shahshahani [J. Appl. Probab. 31A (1994) 49-62]. On the macroscopic scale, $D_N(x)$ gives rise to yet another type of Gaussian process with logarithmic correlations. We give an explicit construction of the latter in terms of a Chebyshev-Fourier random series.
△ Less
Submitted 2 September, 2016; v1 submitted 1 December, 2013;
originally announced December 2013.
-
Tau-Function Theory of Quantum Chaotic Transport with beta=1,2,4
Authors:
F. Mezzadri,
N. J. Simm
Abstract:
We study the cumulants and their generating functions of the probability distributions of the conductance, shot noise and Wigner delay time in ballistic quantum dots. Our approach is based on the integrable theory of certain matrix integrals and applies to all the symmetry classes beta=1,2,4 of Random Matrix Theory. We compute the weak localization corrections to the mixed cumulants of the conduct…
▽ More
We study the cumulants and their generating functions of the probability distributions of the conductance, shot noise and Wigner delay time in ballistic quantum dots. Our approach is based on the integrable theory of certain matrix integrals and applies to all the symmetry classes beta=1,2,4 of Random Matrix Theory. We compute the weak localization corrections to the mixed cumulants of the conductance and shot noise for beta=1,4, thus proving a number of conjectures of Khoruzhenko et al. (Phys. Rev. B, Vol. 80 (2009), 125301). We derive differential equations that characterize the cumulant generating functions for all beta=1,2,4. Furthermore, we show that the cumulant generating function of the Wigner delay time can be expressed in terms of the Painleve' III' transcendant. This allows us to study properties of the cumulants of the Wigner delay time in the asymptotic limit n -> infinity. Finally, for all the symmetry classes and for any number of open channels, we derive a set of recurrence relations that are very efficient for computing cumulants at all orders.
△ Less
Submitted 13 December, 2013; v1 submitted 20 June, 2012;
originally announced June 2012.
-
Moments of the transmission eigenvalues, proper delay times and random matrix theory II
Authors:
F. Mezzadri,
N. J. Simm
Abstract:
We systematically study the first three terms in the asymptotic expansions of the moments of the transmission eigenvalues and proper delay times as the number of quantum channels n in the leads goes to infinity. The computations are based on the assumption that the Landauer-Bütticker scattering matrix for chaotic ballistic cavities can be modelled by the circular ensembles of Random Matrix Theory…
▽ More
We systematically study the first three terms in the asymptotic expansions of the moments of the transmission eigenvalues and proper delay times as the number of quantum channels n in the leads goes to infinity. The computations are based on the assumption that the Landauer-Bütticker scattering matrix for chaotic ballistic cavities can be modelled by the circular ensembles of Random Matrix Theory (RMT). The starting points are the finite-n formulae that we recently discovered (Mezzadri and Simm, J. Math. Phys. 52 (2011), 103511). Our analysis includes all the symmetry classes beta=1,2,4; in addition, it applies to the transmission eigenvalues of Andreev billiards, whose symmetry classes were classified by Zirnbauer (J. Math. Phys. 37 (1996), 4986-5018) and Altland and Zirnbauer (Phys. Rev. B. 55 (1997), 1142-1161). Where applicable, our results are in complete agreement with the semiclassical theory of mesoscopic systems developed by Berkolaiko et al. (J. Phys. A.: Math. Theor. 41 (2008), 365102) and Berkolaiko and Kuipers (J. Phys. A: Math. Theor. 43 (2010), 035101 and New J. Phys. 13 (2011), 063020). Our approach also applies to the Selberg-like integrals. We calculate the first two terms in their asymptotic expansion explicitly.
△ Less
Submitted 10 May, 2012; v1 submitted 14 August, 2011;
originally announced August 2011.
-
Moments of the transmission eigenvalues, proper delay times and random matrix theory I
Authors:
F. Mezzadri,
N. J. Simm
Abstract:
We develop a method to compute the moments of the eigenvalue densities of matrices in the Gaussian, Laguerre and Jacobi ensembles for all the symmetry classes beta = 1,2, 4 and finite matrix dimension n. The moments of the Jacobi ensembles have a physical interpretation as the moments of the transmission eigenvalues of an electron through a quantum dot with chaotic dynamics. For the Laguerre ensem…
▽ More
We develop a method to compute the moments of the eigenvalue densities of matrices in the Gaussian, Laguerre and Jacobi ensembles for all the symmetry classes beta = 1,2, 4 and finite matrix dimension n. The moments of the Jacobi ensembles have a physical interpretation as the moments of the transmission eigenvalues of an electron through a quantum dot with chaotic dynamics. For the Laguerre ensemble we also evaluate the finite n negative moments. Physically, they correspond to the moments of the proper delay times, which are the eigenvalues of the Wigner-Smith matrix. Our formulae are well suited to an asymptotic analysis as n -> infinity.
△ Less
Submitted 23 December, 2011; v1 submitted 31 March, 2011;
originally announced March 2011.
