-
Two-phase training mitigates class imbalance for camera trap image classification with CNNs
Authors:
Farjad Malik,
Simon Wouters,
Ruben Cartuyvels,
Erfan Ghadery,
Marie-Francine Moens
Abstract:
By leveraging deep learning to automatically classify camera trap images, ecologists can monitor biodiversity conservation efforts and the effects of climate change on ecosystems more efficiently. Due to the imbalanced class-distribution of camera trap datasets, current models are biased towards the majority classes. As a result, they obtain good performance for a few majority classes but poor per…
▽ More
By leveraging deep learning to automatically classify camera trap images, ecologists can monitor biodiversity conservation efforts and the effects of climate change on ecosystems more efficiently. Due to the imbalanced class-distribution of camera trap datasets, current models are biased towards the majority classes. As a result, they obtain good performance for a few majority classes but poor performance for many minority classes. We used two-phase training to increase the performance for these minority classes. We trained, next to a baseline model, four models that implemented a different versions of two-phase training on a subset of the highly imbalanced Snapshot Serengeti dataset. Our results suggest that two-phase training can improve performance for many minority classes, with limited loss in performance for the other classes. We find that two-phase training based on majority undersampling increases class-specific F1-scores up to 3.0%. We also find that two-phase training outperforms using only oversampling or undersampling by 6.1% in F1-score on average. Finally, we find that a combination of over- and undersampling leads to a better performance than using them individually.
△ Less
Submitted 29 December, 2021;
originally announced December 2021.
-
Recent developments in the PySCF program package
Authors:
Qiming Sun,
Xing Zhang,
Samragni Banerjee,
Peng Bao,
Marc Barbry,
Nick S. Blunt,
Nikolay A. Bogdanov,
George H. Booth,
Jia Chen,
Zhi-Hao Cui,
Janus Juul Eriksen,
Yang Gao,
Sheng Guo,
Jan Hermann,
Matthew R. Hermes,
Kevin Koh,
Peter Koval,
Susi Lehtola,
Zhendong Li,
Junzi Liu,
Narbe Mardirossian,
James D. McClain,
Mario Motta,
Bastien Mussard,
Hung Q. Pham
, et al. (24 additional authors not shown)
Abstract:
PYSCF is a Python-based general-purpose electronic structure platform that both supports first-principles simulations of molecules and solids, as well as accelerates the development of new methodology and complex computational workflows. The present paper explains the design and philosophy behind PYSCF that enables it to meet these twin objectives. With several case studies, we show how users can…
▽ More
PYSCF is a Python-based general-purpose electronic structure platform that both supports first-principles simulations of molecules and solids, as well as accelerates the development of new methodology and complex computational workflows. The present paper explains the design and philosophy behind PYSCF that enables it to meet these twin objectives. With several case studies, we show how users can easily implement their own methods using PYSCF as a development environment. We then summarize the capabilities of PYSCF for molecular and solid-state simulations. Finally, we describe the growing ecosystem of projects that use PYSCF across the domains of quantum chemistry, materials science, machine learning and quantum information science.
△ Less
Submitted 10 July, 2020; v1 submitted 27 February, 2020;
originally announced February 2020.
-
T3NS: three-legged tree tensor network states
Authors:
Klaas Gunst,
Frank Verstraete,
Sebastian Wouters,
Örs Legeza,
Dimitri Van Neck
Abstract:
We present a new variational tree tensor network state (TTNS) ansatz, the three-legged tree tensor network state (T3NS). Physical tensors are interspersed with branching tensors. Physical tensors have one physical index and at most two virtual indices, as in the matrix product state (MPS) ansatz of the density matrix renormalization group (DMRG). Branching tensors have no physical index, but up to…
▽ More
We present a new variational tree tensor network state (TTNS) ansatz, the three-legged tree tensor network state (T3NS). Physical tensors are interspersed with branching tensors. Physical tensors have one physical index and at most two virtual indices, as in the matrix product state (MPS) ansatz of the density matrix renormalization group (DMRG). Branching tensors have no physical index, but up to three virtual indices. In this way, advantages of DMRG, in particular a low computational cost and a simple implementation of symmetries, are combined with advantages of TTNS, namely incorporating more entanglement. Our code is capable of simulating quantum chemical Hamiltonians, and we present several proof-of-principle calculations on LiF, N$_2$ and the bis($μ$-oxo) and $μ- η^2 : η^2$ peroxo isomers of $[\mathrm{Cu}_2\mathrm{O}_2]^{2+}$.
△ Less
Submitted 30 January, 2018;
originally announced January 2018.
-
Measurements of the energy distribution of an ultracold rubidium ion beam
Authors:
G. ten Haaf,
S. H. W. Wouters,
D. F. J. Nijhof,
P. H. A. Mutsaers,
E. J. D. Vredenbregt
Abstract:
The energy distribution of an ultracold rubidium ion beam, which is intended to be used as the source for a focused ion beam instrument, is measured with a retarding field analyzer. The ions are created from a laser-cooled and compressed atomic beam by two-step photoionization in which the ionization laser power is enhanced in a build-up cavity. Particle tracing simulations are performed to ensure…
▽ More
The energy distribution of an ultracold rubidium ion beam, which is intended to be used as the source for a focused ion beam instrument, is measured with a retarding field analyzer. The ions are created from a laser-cooled and compressed atomic beam by two-step photoionization in which the ionization laser power is enhanced in a build-up cavity. Particle tracing simulations are performed to ensure the analyzer is able to resolve the distribution. The lowest achieved full width 50% energy spread is $\left(0.205\pm0.006\right)$ eV. The energy spread originates from the variation in the ionization position of the ions which are created inside an extraction electric field. This extraction field is essential to limit disorder-induced heating which can decrease the ion beam brightness. The ionization position distribution is limited by a tightly focused excitation laser beam. Energy distributions are measured for various ionization and excitation laser intensities and compared with calculations based on numerical solutions of the optical Bloch equations including ionization. A good agreement is found between measurements and calculations.
△ Less
Submitted 31 October, 2017;
originally announced October 2017.
-
Cavity-enhanced photoionization of an ultracold rubidium beam for application in focused ion beams
Authors:
G. ten Haaf,
S. H. W. Wouters,
P. H. A. Mutsaers,
E. J. D. Vredenbregt
Abstract:
A two-step photoionization strategy of an ultracold rubidium beam for application in a focused ion beam instrument is analyzed and implemented. In this strategy the atomic beam is partly selected with an aperture after which the transmitted atoms are ionized in the overlap of a tightly cylindrically focused excitation laser beam and an ionization laser beam whose power is enhanced in a build-up ca…
▽ More
A two-step photoionization strategy of an ultracold rubidium beam for application in a focused ion beam instrument is analyzed and implemented. In this strategy the atomic beam is partly selected with an aperture after which the transmitted atoms are ionized in the overlap of a tightly cylindrically focused excitation laser beam and an ionization laser beam whose power is enhanced in a build-up cavity. The advantage of this strategy, as compared to without the use of a build-up cavity, is that higher ionization degrees can be reached at higher currents. Optical Bloch equations including the photoionization process are used to calculate what ionization degree and ionization position distribution can be reached. Furthermore, the ionization strategy is tested on an ultracold beam of $^{85}$Rb atoms. The beam current is measured as a function of the excitation and ionization laser beam intensity and the selection aperture size. Although details are different, the global trends of the measurements agree well with the calculation. With a selection aperture diameter of 52 $μ$m, a current of $\left(170\pm4\right)$ pA is measured, which according to calculations is 63% of the current equivalent of the transmitted atomic flux. Taking into account the ionization degree the ion beam peak reduced brightness is estimated at $1\times10^7$ A/(m$^2\,$sr$\,$eV).
△ Less
Submitted 31 October, 2017; v1 submitted 6 August, 2017;
originally announced August 2017.
-
Block product density matrix embedding theory for strongly correlated spin systems
Authors:
Klaas Gunst,
Sebastian Wouters,
Stijn De Baerdemacker,
Dimitri Van Neck
Abstract:
Density matrix embedding theory (DMET) is a relatively new technique for the calculation of strongly correlated systems. Recently, block product DMET (BPDMET) was introduced for the study of spin systems such as the antiferromagnetic $J_1 - J_2$ model on the square lattice. In this paper, we extend the variational Ansatz of BPDMET using spin-state optimization, yielding improved results. We apply…
▽ More
Density matrix embedding theory (DMET) is a relatively new technique for the calculation of strongly correlated systems. Recently, block product DMET (BPDMET) was introduced for the study of spin systems such as the antiferromagnetic $J_1 - J_2$ model on the square lattice. In this paper, we extend the variational Ansatz of BPDMET using spin-state optimization, yielding improved results. We apply the same techniques to the Kitaev-Heisenberg model on the honeycomb lattice, comparing the results when using several types of clusters. Energy profiles and correlation functions are investigated. A diagonalization in the tangent space of the variational approach yields information on the excited states and the corresponding spectral functions.
△ Less
Submitted 29 May, 2017; v1 submitted 14 February, 2017;
originally announced February 2017.
-
The Python-based Simulations of Chemistry Framework (PySCF)
Authors:
Qiming Sun,
Timothy C. Berkelbach,
Nick S. Blunt,
George H. Booth,
Sheng Guo,
Zhendong Li,
Junzi Liu,
James McClain,
Elvira R. Sayfutyarova,
Sandeep Sharma,
Sebastian Wouters,
Garnet Kin-Lic Chan
Abstract:
PySCF is a general-purpose electronic structure platform designed from the ground up to emphasize code simplicity, both to aid new method development, as well as for flexibility in computational workflow. The package provides a wide range of tools to support simulations of finite size systems, extended systems with periodic boundary conditions, low dimensional periodic systems, and custom Hamilton…
▽ More
PySCF is a general-purpose electronic structure platform designed from the ground up to emphasize code simplicity, both to aid new method development, as well as for flexibility in computational workflow. The package provides a wide range of tools to support simulations of finite size systems, extended systems with periodic boundary conditions, low dimensional periodic systems, and custom Hamiltonians, using mean-field and post-mean-field methods with standard Gaussian basis functions. To ensure easy of extensibility, PySCF uses the Python language to implement almost all its features, while computationally critical paths are implemented with heavily optimized C routines. Using this combined Python/C implementation, the package is as efficient as the best existing C or Fortran based quantum chemistry programs. In this paper we document the capabilities and design philosophy of the current version of the PySCF package.
△ Less
Submitted 2 March, 2017; v1 submitted 27 January, 2017;
originally announced January 2017.
-
Direct magneto-optical compression of an effusive atomic beam for high-resolution focused ion beam application
Authors:
G. ten Haaf,
T. C. H. de Raadt,
G. P. Offermans,
J. F. M. van Rens,
P. H. A. Mutsaers,
E. J. D. Vredenbregt,
S. H. W. Wouters
Abstract:
An atomic rubidium beam formed in a 70 mm long two-dimensional magneto-optical trap (2D MOT), directly loaded from a collimated Knudsen source, is analyzed using laser-induced fluorescence. The longitudinal velocity distribution, the transverse temperature and the flux of the atomic beam are reported. The equivalent transverse reduced brightness of an ion beam with similar properties as the atomic…
▽ More
An atomic rubidium beam formed in a 70 mm long two-dimensional magneto-optical trap (2D MOT), directly loaded from a collimated Knudsen source, is analyzed using laser-induced fluorescence. The longitudinal velocity distribution, the transverse temperature and the flux of the atomic beam are reported. The equivalent transverse reduced brightness of an ion beam with similar properties as the atomic beam is calculated because the beam is developed to be photoionized and applied in a focused ion beam. In a single two-dimensional magneto-optical trap** step an equivalent transverse reduced brightness of $(1.0\substack{+0.8-0.4})$ $\times 10^6$ A/(m$^2$ sr eV) was achieved with a beam flux equivalent to $(0.6\substack{+0.3-0.2})$ nA. The temperature of the beam is further reduced with an optical molasses after the 2D MOT. This increased the equivalent brightness to $(6\substack{+5-2})$$\times 10^6$ A/(m$^2$ sr eV). For currents below 10 pA, for which disorder-induced heating can be suppressed, this number is also a good estimate of the ion beam brightness that can be expected. Such an ion beam brightness would be a six times improvement over the liquid metal ion source and could improve the resolution in focused ion beam nanofabrication.
△ Less
Submitted 18 April, 2017; v1 submitted 6 December, 2016;
originally announced December 2016.
-
Five years of density matrix embedding theory
Authors:
Sebastian Wouters,
Carlos A. Jiménez-Hoyos,
Garnet K. -L. Chan
Abstract:
Density matrix embedding theory (DMET) describes finite fragments in the presence of a surrounding environment. In contrast to most embedding methods, DMET explicitly allows for quantum entanglement between both. In this chapter, we discuss both the ground-state and response theory formulations of DMET, and review several applications. In addition, a proof is given that the local density of states…
▽ More
Density matrix embedding theory (DMET) describes finite fragments in the presence of a surrounding environment. In contrast to most embedding methods, DMET explicitly allows for quantum entanglement between both. In this chapter, we discuss both the ground-state and response theory formulations of DMET, and review several applications. In addition, a proof is given that the local density of states can be obtained by working with a Fock space of bath orbitals.
△ Less
Submitted 31 March, 2016;
originally announced May 2016.
-
DMRG-CASPT2 study of the longitudinal static second hyperpolarizability of all-trans polyenes
Authors:
Sebastian Wouters,
Veronique Van Speybroeck,
Dimitri Van Neck
Abstract:
We have implemented internally contracted complete active space second order perturbation theory (CASPT2) with the density matrix renormalization group (DMRG) as active space solver [Y. Kurashige and T. Yanai, J. Chem. Phys. 135, 094104 (2011)]. Internally contracted CASPT2 requires to contract the generalized Fock matrix with the 4-particle reduced density matrix (4-RDM) of the reference wavefunc…
▽ More
We have implemented internally contracted complete active space second order perturbation theory (CASPT2) with the density matrix renormalization group (DMRG) as active space solver [Y. Kurashige and T. Yanai, J. Chem. Phys. 135, 094104 (2011)]. Internally contracted CASPT2 requires to contract the generalized Fock matrix with the 4-particle reduced density matrix (4-RDM) of the reference wavefunction. The required 4-RDM elements can be obtained from 3-particle reduced density matrices (3-RDM) of different wavefunctions, formed by symmetry-conserving single-particle excitations op top of the reference wavefunction. In our spin-adapted DMRG code chemps2 [https://github.com/sebwouters/chemps2], we decompose these excited wavefunctions as spin-adapted matrix product states, and calculate their 3-RDM in order to obtain the required contraction of the generalized Fock matrix with the 4-RDM of the reference wavefunction. In this work, we study the longitudinal static second hyperpolarizability of all-trans polyenes C$_{2n}$H$_{2n+2}$ [n = 4 - 12] in the cc-pVDZ basis set. DMRG-SCF and DMRG-CASPT2 yield substantially lower values and scaling with system size compared to RHF and MP2, respectively.
△ Less
Submitted 13 July, 2016; v1 submitted 18 May, 2016;
originally announced May 2016.
-
A practical guide to density matrix embedding theory in quantum chemistry
Authors:
Sebastian Wouters,
Carlos A. Jiménez-Hoyos,
Qiming Sun,
Garnet Kin-Lic Chan
Abstract:
Density matrix embedding theory (DMET) provides a theoretical framework to treat finite fragments in the presence of a surrounding molecular or bulk environment, even when there is significant correlation or entanglement between the two. In this work, we give a practically oriented and explicit description of the numerical and theoretical formulation of DMET. We also describe in detail how to perf…
▽ More
Density matrix embedding theory (DMET) provides a theoretical framework to treat finite fragments in the presence of a surrounding molecular or bulk environment, even when there is significant correlation or entanglement between the two. In this work, we give a practically oriented and explicit description of the numerical and theoretical formulation of DMET. We also describe in detail how to perform self-consistent DMET optimizations. We explore different embedding strategies with and without a self-consistency condition in hydrogen rings, beryllium rings, and a sample S$_{\text{N}}$2 reaction. The source code for the calculations in this work can be obtained from \url{https://github.com/sebwouters/qc-dmet}.
△ Less
Submitted 5 May, 2016; v1 submitted 28 March, 2016;
originally announced March 2016.
-
Design and experimental validation of a compact collimated Knudsen source
Authors:
Steinar H W Wouters,
Gijs ten Haaf,
Peter H A Mutsaers,
Edgar J D Vredenbregt
Abstract:
In this paper we discuss the design and performance of a collimated Knudsen source which has the benefit of a simple design over recirculating sources. Measurements of the flux, transverse velocity distribution and brightness at different temperatures were conducted to evaluate the performance. The scaling of the flux and brightness with the source temperature follow the theoretical predictions. T…
▽ More
In this paper we discuss the design and performance of a collimated Knudsen source which has the benefit of a simple design over recirculating sources. Measurements of the flux, transverse velocity distribution and brightness at different temperatures were conducted to evaluate the performance. The scaling of the flux and brightness with the source temperature follow the theoretical predictions. The transverse velocity distribution in the transparent operation regime also agrees with the simulated data. The source was found able to produce a flux of $10^{14}$ s$^{-1}$ at a temperature of 433 K. Furthermore the transverse reduced brightness of an ion beam with equal properties as the atomic beam reads $1.7 \times 10^2$ A/(m${}^2$ sr eV) which is sufficient for our goal: the creation of an ultra-cold ion beam by ionization of a laser-cooled and compressed atomic rubidium beam.
△ Less
Submitted 24 November, 2015;
originally announced November 2015.
-
Performance predictions of a focused ion beam from a laser cooled and compressed atomic beam
Authors:
G. ten Haaf,
S. H. W. Wouters,
S. B. van der Geer,
E. J. D. Vredenbregt,
P. H. A. Mutsaers
Abstract:
Focused ion beams are indispensable tools in the semiconductor industry because of their ability to image and modify structures at the nanometer length scale. Here we report on performance predictions of a new type of focused ion beam based on photo-ionization of a laser cooled and compressed atomic beam. Particle tracing simulations are performed to investigate the effects of disorder-induced hea…
▽ More
Focused ion beams are indispensable tools in the semiconductor industry because of their ability to image and modify structures at the nanometer length scale. Here we report on performance predictions of a new type of focused ion beam based on photo-ionization of a laser cooled and compressed atomic beam. Particle tracing simulations are performed to investigate the effects of disorder-induced heating after ionization in a large electric field. They lead to a constraint on this electric field strength which is used as input for an analytical model which predicts the minimum attainable spot size as a function of amongst others the flux density of the atomic beam, the temperature of this beam and the total current. At low currents (I<10 pA) the spot size will be limited by a combination of spherical aberration and brightness, while at higher currents this is a combination of chromatic aberration and brightness. It is expected that a nanometer size spot is possible at a current of 1 pA. The analytical model was verified with particle tracing simulations of a complete focused ion beam setup. A genetic algorithm was used to find the optimum acceleration electric field as a function of the current. At low currents the result agrees well with the analytical model while at higher currents the spot sizes found are even lower due to effects that are not taken into account in the analytical model.
△ Less
Submitted 26 January, 2015; v1 submitted 16 October, 2014;
originally announced October 2014.
-
Performance predictions for a laser intensified thermal beam for use in high resolution Focused Ion Beam instruments
Authors:
S. H. W. Wouters,
G. ten Haaf,
R. P. M. J. W. Notermans,
N. Debernardi,
P. H. A. Mutsaers,
O. J. Luiten,
E. J. D. Vredenbregt
Abstract:
Photo-ionization of a laser-cooled and compressed atomic beam from a high-flux thermal source can be used to create a high-brightness ion beam for use in Focus Ion Beam (FIB) instruments. Here we show using calculations and Doppler cooling simulations that an atomic rubidium beam with a brightness of $2.1 \times 10^7 A/(m^2\,sr\,eV)$ at a current of 1 nA can be created using a compact 5 cm long 2D…
▽ More
Photo-ionization of a laser-cooled and compressed atomic beam from a high-flux thermal source can be used to create a high-brightness ion beam for use in Focus Ion Beam (FIB) instruments. Here we show using calculations and Doppler cooling simulations that an atomic rubidium beam with a brightness of $2.1 \times 10^7 A/(m^2\,sr\,eV)$ at a current of 1 nA can be created using a compact 5 cm long 2D magneto-optical compressor which is more than an order of magnitude better than the current state of the art Liquid Metal Ion Source.
△ Less
Submitted 17 October, 2014; v1 submitted 16 October, 2014;
originally announced October 2014.
-
The density matrix renormalization group for ab initio quantum chemistry
Authors:
Sebastian Wouters,
Dimitri Van Neck
Abstract:
During the past 15 years, the density matrix renormalization group (DMRG) has become increasingly important for ab initio quantum chemistry. Its underlying wavefunction ansatz, the matrix product state (MPS), is a low-rank decomposition of the full configuration interaction tensor. The virtual dimension of the MPS, the rank of the decomposition, controls the size of the corner of the many-body Hil…
▽ More
During the past 15 years, the density matrix renormalization group (DMRG) has become increasingly important for ab initio quantum chemistry. Its underlying wavefunction ansatz, the matrix product state (MPS), is a low-rank decomposition of the full configuration interaction tensor. The virtual dimension of the MPS, the rank of the decomposition, controls the size of the corner of the many-body Hilbert space that can be reached with the ansatz. This parameter can be systematically increased until numerical convergence is reached. The MPS ansatz naturally captures exponentially decaying correlation functions. Therefore DMRG works extremely well for noncritical one-dimensional systems. The active orbital spaces in quantum chemistry are however often far from one-dimensional, and relatively large virtual dimensions are required to use DMRG for ab initio quantum chemistry (QC-DMRG). The QC-DMRG algorithm, its computational cost, and its properties are discussed. Two important aspects to reduce the computational cost are given special attention: the orbital choice and ordering, and the exploitation of the symmetry group of the Hamiltonian. With these considerations, the QC-DMRG algorithm allows to find numerically exact solutions in active spaces of up to 40 electrons in 40 orbitals.
△ Less
Submitted 8 July, 2014;
originally announced July 2014.
-
DMRG-SCF study of the singlet, triplet, and quintet states of oxo-Mn(Salen)
Authors:
Sebastian Wouters,
Thomas Bogaerts,
Pascal Van Der Voort,
Veronique Van Speybroeck,
Dimitri Van Neck
Abstract:
We use CheMPS2, our free open-source spin-adapted implementation of the density matrix renormalization group (DMRG) [Wouters et al., Comput. Phys. Commun. 185, 1501 (2014)], to study the lowest singlet, triplet, and quintet states of the oxo-Mn(Salen) complex. We describe how an initial approximate DMRG calculation in a large active space around the Fermi level can be used to obtain a good set of…
▽ More
We use CheMPS2, our free open-source spin-adapted implementation of the density matrix renormalization group (DMRG) [Wouters et al., Comput. Phys. Commun. 185, 1501 (2014)], to study the lowest singlet, triplet, and quintet states of the oxo-Mn(Salen) complex. We describe how an initial approximate DMRG calculation in a large active space around the Fermi level can be used to obtain a good set of starting orbitals for subsequent complete-active-space or DMRG self-consistent field (CASSCF or DMRG-SCF) calculations. This procedure mitigates the need for a localization procedure, followed by a manual selection of the active space. Per multiplicity, the same active space of 28 electrons in 22 orbitals (28e, 22o) is obtained with the 6-31G*, cc-pVDZ, and ANO-RCC-VDZP basis sets (the latter with DKH2 scalar relativistic corrections). Our calculations provide new insight into the electronic structure of the quintet.
△ Less
Submitted 18 June, 2014; v1 submitted 22 May, 2014;
originally announced May 2014.
-
Accurate variational electronic structure calculations with the density matrix renormalization group
Authors:
Sebastian Wouters
Abstract:
During the past 15 years, the density matrix renormalization group (DMRG) has become increasingly important for ab initio quantum chemistry. The underlying matrix product state (MPS) ansatz is a low-rank decomposition of the full configuration interaction tensor. The virtual dimension of the MPS controls the size of the corner of the many-body Hilbert space that can be reached.
Whereas the MPS a…
▽ More
During the past 15 years, the density matrix renormalization group (DMRG) has become increasingly important for ab initio quantum chemistry. The underlying matrix product state (MPS) ansatz is a low-rank decomposition of the full configuration interaction tensor. The virtual dimension of the MPS controls the size of the corner of the many-body Hilbert space that can be reached.
Whereas the MPS ansatz will only yield an efficient description for noncritical one-dimensional systems, it can still be used as a variational ansatz for other finite-size systems. Rather large virtual dimensions are then required. The two most important aspects to reduce the corresponding computational cost are a proper choice and ordering of the active space orbitals, and the exploitation of the symmetry group of the Hamiltonian. By taking care of both aspects, DMRG becomes an efficient replacement for exact diagonalization in quantum chemistry.
DMRG and Hartree-Fock theory have an analogous structure. The former can be interpreted as a self-consistent mean-field theory in the DMRG lattice sites, and the latter in the particles. It is possible to build upon this analogy to introduce post-DMRG methods. Based on an approximate MPS, these methods provide improved ansätze for the ground state, as well as for excitations. Exponentiation of the single-particle (single-site) excitations for a Slater determinant (an MPS with open boundary conditions) leads to the Thouless theorem for Hartree-Fock theory (DMRG), an explicit nonredundant parameterization of the entire manifold of Slater determinants (MPS wavefunctions). This gives rise to the configuration interaction expansion for DMRG. The Hubbard-Stratonovich transformation lies at the basis of auxiliary field quantum Monte Carlo for Slater determinants. An analogous transformation for spin-lattice Hamiltonians allows to formulate a promising variant for MPSs.
△ Less
Submitted 6 May, 2014;
originally announced May 2014.
-
Projector quantum Monte Carlo with matrix product states
Authors:
Sebastian Wouters,
Brecht Verstichel,
Dimitri Van Neck,
Garnet Kin-Lic Chan
Abstract:
We marry tensor network states (TNS) and projector quantum Monte Carlo (PMC) to overcome the high computational scaling of TNS and the sign problem of PMC. Using TNS as trial wavefunctions provides a route to systematically improve the sign structure and to eliminate the bias in fixed-node and constrained-path PMC. As a specific example, we describe phaseless auxiliary-field quantum Monte Carlo wi…
▽ More
We marry tensor network states (TNS) and projector quantum Monte Carlo (PMC) to overcome the high computational scaling of TNS and the sign problem of PMC. Using TNS as trial wavefunctions provides a route to systematically improve the sign structure and to eliminate the bias in fixed-node and constrained-path PMC. As a specific example, we describe phaseless auxiliary-field quantum Monte Carlo with matrix product states (MPS-AFQMC). MPS-AFQMC improves significantly on the DMRG ground-state energy. For the J1-J2 model on two-dimensional square lattices, we observe with MPS-AFQMC an order of magnitude reduction in the error for all couplings, compared to DMRG. The improvement is independent of walker bond dimension, and we therefore use bond dimension one for the walkers. The computational cost of MPS-AFQMC is then quadratic in the bond dimension of the trial wavefunction, which is lower than the cubic scaling of DMRG. The error due to the constrained-path bias is proportional to the variational error of the trial wavefunction. We show that for the J1-J2 model on two-dimensional square lattices, a linear extrapolation of the MPS-AFQMC energy with the discarded weight from the DMRG calculation allows to remove the constrained-path bias. Extensions to other tensor networks are briefly discussed.
△ Less
Submitted 14 May, 2014; v1 submitted 12 March, 2014;
originally announced March 2014.
-
CheMPS2: a free open-source spin-adapted implementation of the density matrix renormalization group for ab initio quantum chemistry
Authors:
Sebastian Wouters,
Ward Poelmans,
Paul W. Ayers,
Dimitri Van Neck
Abstract:
The density matrix renormalization group (DMRG) has become an indispensable numerical tool to find exact eigenstates of finite-size quantum systems with strong correlation. In the fields of condensed matter, nuclear structure and molecular electronic structure, it has significantly extended the system sizes that can be handled compared to full configuration interaction, without losing numerical ac…
▽ More
The density matrix renormalization group (DMRG) has become an indispensable numerical tool to find exact eigenstates of finite-size quantum systems with strong correlation. In the fields of condensed matter, nuclear structure and molecular electronic structure, it has significantly extended the system sizes that can be handled compared to full configuration interaction, without losing numerical accuracy. For quantum chemistry (QC), the most efficient implementations of DMRG require the incorporation of particle number, spin and point group symmetries in the underlying matrix product state (MPS) ansatz, as well as the use of so-called complementary operators. The symmetries introduce a sparse block structure in the MPS ansatz and in the intermediary contracted tensors. If a symmetry is non-abelian, the Wigner-Eckart theorem allows to factorize a tensor into a Clebsch-Gordan coefficient and a reduced tensor. In addition, the fermion signs have to be carefully tracked. Because of these challenges, implementing DMRG efficiently for QC is not straightforward. Efficient and freely available implementations are therefore highly desired. In this work we present CheMPS2, our free open-source spin-adapted implementation of DMRG for ab initio QC. Around CheMPS2, we have implemented the augmented Hessian Newton-Raphson complete active space self-consistent field method, with exact Hessian. The bond dissociation curves of the 12 lowest states of the carbon dimer were obtained at the DMRG(28 orbitals, 12 electrons, D$_{\mathsf{SU(2)}}$=2500)/cc-pVDZ level of theory. The contribution of $1s$ core correlation to the $X^1Σ_g^+$ bond dissociation curve of the carbon dimer was estimated by comparing energies at the DMRG(36o, 12e, D$_{\mathsf{SU(2)}}$=2500)/cc-pCVDZ and DMRG-SCF(34o, 8e, D$_{\mathsf{SU(2)}}$=2500)/cc-pCVDZ levels of theory.
△ Less
Submitted 8 January, 2014; v1 submitted 9 December, 2013;
originally announced December 2013.
-
Linear Response Theory for the Density Matrix Renormalization Group: Efficient Algorithms for Strongly Correlated Excited States
Authors:
Naoki Nakatani,
Sebastian Wouters,
Dimitri Van Neck,
Garnet Kin-Lic Chan
Abstract:
Linear response theory for the density matrix renormalization group (DMRG-LRT) was first presented in terms of the DMRG renormalization projectors [Dorando et al., J. Chem. Phys. 130, 184111 (2009)]. Later, with an understanding of the manifold structure of the matrix product state (MPS) ansatz, which lies at the basis of the DMRG algorithm, a way was found to construct the linear response space f…
▽ More
Linear response theory for the density matrix renormalization group (DMRG-LRT) was first presented in terms of the DMRG renormalization projectors [Dorando et al., J. Chem. Phys. 130, 184111 (2009)]. Later, with an understanding of the manifold structure of the matrix product state (MPS) ansatz, which lies at the basis of the DMRG algorithm, a way was found to construct the linear response space for general choices of the MPS gauge in terms of the tangent space vectors [Haegeman et al., Phys. Rev. Lett. 107, 070601 (2011)]. These two developments led to the formulation of the Tamm-Dancoff and random phase approximations (TDA and RPA) for MPS. This work describes how these LRTs may be efficiently implemented through minor modifications of the DMRG sweep algorithm, at a computational cost which scales the same as the ground-state DMRG algorithm. In fact, the mixed canonical MPS form implicit to the DMRG sweep is essential for efficient implementation of the RPA, due to the structure of the second-order tangent space. We present ab initio DMRG-TDA results for excited states of polyenes, the water molecule, and a [2Fe-2S] iron-sulfur cluster.
△ Less
Submitted 7 November, 2013;
originally announced November 2013.
-
Variational optimization of the 2DM: approaching three-index accuracy using extended cluster constraints
Authors:
Brecht Verstichel,
Ward Poelmans,
Stijn De Baerdemacker,
Sebastian Wouters,
Dimitri Van Neck
Abstract:
The reduced density matrix is variationally optimized for the two-dimensional Hubbard model. Exploiting all symmetries present in the system, we have been able to study $6\times6$ lattices at various fillings and different values for the on-site repulsion, using the highly accurate but computationally expensive three-index conditions. To reduce the computational cost we study the performance of im…
▽ More
The reduced density matrix is variationally optimized for the two-dimensional Hubbard model. Exploiting all symmetries present in the system, we have been able to study $6\times6$ lattices at various fillings and different values for the on-site repulsion, using the highly accurate but computationally expensive three-index conditions. To reduce the computational cost we study the performance of imposing the three-index constraints on local clusters of $2\times2$ and $3\times3$ sites. We subsequently derive new constraints which extend these cluster constraints to incorporate the open-system nature of a cluster on a larger lattice. The feasibility of implementing these new constraints is demonstrated by performing a proof-of-principle calculation on the $6\times6$ lattice. It is shown that a large portion of the three-index result can be recovered using these extended cluster constraints, at a fraction of the computational cost.
△ Less
Submitted 13 March, 2014; v1 submitted 3 July, 2013;
originally announced July 2013.
-
The Thouless theorem for matrix product states and subsequent post-density matrix renormalization group methods
Authors:
Sebastian Wouters,
Naoki Nakatani,
Dimitri Van Neck,
Garnet Kin-Lic Chan
Abstract:
The similarities between Hartree-Fock (HF) theory and the density-matrix renormalization group (DMRG) are explored. Both methods can be formulated as the variational optimization of a wave-function ansatz. Linearization of the time-dependent variational principle near a variational minimum allows to derive the random phase approximation (RPA). We show that the non-redundant parametrization of the…
▽ More
The similarities between Hartree-Fock (HF) theory and the density-matrix renormalization group (DMRG) are explored. Both methods can be formulated as the variational optimization of a wave-function ansatz. Linearization of the time-dependent variational principle near a variational minimum allows to derive the random phase approximation (RPA). We show that the non-redundant parametrization of the matrix product state (MPS) tangent space [J. Haegeman et al., Phys. Rev. Lett. 107, 070601 (2011)] leads to the Thouless theorem for MPS, i.e. an explicit non-redundant parametrization of the entire MPS manifold, starting from a specific MPS reference. Excitation operators are identified, which extends the analogy between HF and DMRG to the Tamm-Dancoff approximation (TDA), the configuration interaction (CI) expansion, and coupled cluster theory. For a small one-dimensional Hubbard chain, we use a CI-MPS ansatz with single and double excitations to improve on the ground state and to calculate low-lying excitation energies. For a symmetry-broken ground state of this model, we show that RPA-MPS allows to retrieve the Goldstone mode. We also discuss calculations of the RPA-MPS correlation energy. With the long-range quantum chemical Pariser-Parr-Pople Hamiltonian, low-lying TDA-MPS and RPA-MPS excitation energies for polyenes are obtained.
△ Less
Submitted 19 July, 2013; v1 submitted 8 May, 2013;
originally announced May 2013.
-
Extensive v2DM study of the one-dimensional Hubbard model for large lattice sizes: Exploiting translational invariance and parity
Authors:
Brecht Verstichel,
Helen van Aggelen,
Ward Poelmans,
Sebastian Wouters,
Dimitri Van Neck
Abstract:
Using variational density matrix optimization with two- and three-index conditions we study the one-dimensional Hubbard model with periodic boundary conditions at various filling factors. Special attention is directed to the full exploitation of the available symmetries, more specifically the combination of translational invariance and space-inversion parity, which allows for the study of large la…
▽ More
Using variational density matrix optimization with two- and three-index conditions we study the one-dimensional Hubbard model with periodic boundary conditions at various filling factors. Special attention is directed to the full exploitation of the available symmetries, more specifically the combination of translational invariance and space-inversion parity, which allows for the study of large lattice sizes. We compare the computational scaling of three different semidefinite programming algorithms with increasing lattice size, and find the boundary point method to be the most suited for this type of problem. Several physical properties, such as the two-particle correlation functions, are extracted to check the physical content of the variationally determined density matrix. It is found that the three-index conditions are needed to correctly describe the full phase diagram of the Hubbard model. We also show that even in the case of half filling, where the ground-state energy is close to the exact value, other properties such as the spin-correlation function can be flawed.
△ Less
Submitted 1 March, 2013;
originally announced March 2013.
-
Longitudinal static optical properties of hydrogen chains: finite field extrapolations of matrix product state calculations
Authors:
Sebastian Wouters,
Peter A. Limacher,
Dimitri Van Neck,
Paul W. Ayers
Abstract:
We have implemented the sweep algorithm for the variational optimization of SU(2) x U(1) (spin and particle number) invariant matrix product states (MPS) for general spin and particle number invariant fermionic Hamiltonians. This class includes non-relativistic quantum chemical systems within the Born-Oppenheimer approximation. High-accuracy ab-initio finite field results of the longitudinal stati…
▽ More
We have implemented the sweep algorithm for the variational optimization of SU(2) x U(1) (spin and particle number) invariant matrix product states (MPS) for general spin and particle number invariant fermionic Hamiltonians. This class includes non-relativistic quantum chemical systems within the Born-Oppenheimer approximation. High-accuracy ab-initio finite field results of the longitudinal static polarizabilities and second hyperpolarizabilities of one-dimensional hydrogen chains are presented. This allows to assess the performance of other quantum chemical methods. For small basis sets, MPS calculations in the saturation regime of the optical response properties can be performed. These results are extrapolated to the thermodynamic limit.
△ Less
Submitted 15 March, 2012; v1 submitted 1 February, 2012;
originally announced February 2012.
-
High-accuracy ab-initio quantum chemistry by means of an SU(2) x U(1) invariant matrix product state Ansatz: the static second hyperpolarizability
Authors:
Sebastian Wouters,
Peter A. Limacher,
Dimitri Van Neck,
Paul W. Ayers
Abstract:
We have implemented the single-site density matrix renormalization group algorithm for the variational optimization of SU(2) \times U(1) (spin and particle number) invariant matrix product states for general spin and particle number symmetric fermionic Hamiltonians. This class also includes non-relativistic quantum chemical systems within the Born-Oppenheimer approximation. High-accuracy ab-initio…
▽ More
We have implemented the single-site density matrix renormalization group algorithm for the variational optimization of SU(2) \times U(1) (spin and particle number) invariant matrix product states for general spin and particle number symmetric fermionic Hamiltonians. This class also includes non-relativistic quantum chemical systems within the Born-Oppenheimer approximation. High-accuracy ab-initio finite field results of the longitudinal static polarizabilities and second hyperpolarizabilities of one-dimensional hydrogen chains are obtained with the algorithm. A comparison with other methods is made.
△ Less
Submitted 24 November, 2011;
originally announced November 2011.
