arXiv subjects

Chang-Yu Hsieh

Publications and source records attributed to Chang-Yu Hsieh.

At least 73 records · Page 4Linked to original sources

arXiv2019

Alchemy: A Quantum Chemistry Dataset for Benchmarking AI Models

We introduce a new molecular dataset, named Alchemy, for developing machine learning models useful in chemistry and material science. As of June 20th 2019, the dataset comprises of 12 quantum mechanical properties of 119,487 organic molecules with up to 14 heavy atoms, sampled from the GDB MedChem database. The Alchemy dataset expands the volume and diversity of existing molecular datasets. Our extensive benchmarks of the state-of-the-art graph neural network models on Alchemy clearly manifest the usefulness of new data in validating and developing machine learning models for chemistry and material science. We further launch a contest to attract attentions from researchers in the related fields. More details can be found on the contest website \footnote{https://alchemy.tencent.com}. At the time of benchamrking experiment, we have generated 119,487 molecules in our Alchemy dataset. More molecular samples are generated since then. Hence, we provide a list of molecules used in the reported benchmarks.

cs.LG↗

arXiv2019

Utilizing Edge Features in Graph Neural Networks via Variational Information Maximization

Graph Neural Networks (GNNs) achieve an impressive performance on structured graphs by recursively updating the representation vector of each node based on its neighbors, during which parameterized transformation matrices should be learned for the node feature updating. However, existing propagation schemes are far from being optimal since they do not fully utilize the relational information between nodes. We propose the information maximizing graph neural networks (IGNN), which maximizes the mutual information between edge states and transform parameters. We reformulate the mutual information as a differentiable objective via a variational approach. We compare our model against several recent variants of GNNs and show that our model achieves the state-of-the-art performance on multiple tasks including quantum chemistry regression on QM9 dataset, generalization capability from QM9 to larger molecular graphs, and prediction of molecular bioactivities relevant for drug discovery. The IGNN model is based on an elegant and fundamental idea in information theory as explained in the main text, and it could be easily generalized beyond the contexts of molecular graphs considered in this work. To encourage more future work in this area, all datasets and codes used in this paper will be released for public access.

cs.LG↗

arXiv2019

Non-Markovian Noise Characterization with the Transfer Tensor Method

We propose simple protocols for performing quantum noise spectroscopy based on the method of transfer tensor maps (TTM), [Phys. Rev. Lett. 112, 110401 (2014)]. The TTM approach is a systematic way to deduce the memory kernel of a time-nonlocal quantum master equation via quantum process tomography. With access to the memory kernel it is possible to (1) assess the non-Markovianity of a quantum process, (2) reconstruct the noise spectral density beyond pure dephasing models, and (3) investigate collective decoherence in multiqubit devices. We illustrate the usefulness of TTM spectroscopy on the IBM Quantum Experience platform, and demonstrate that the qubits in the IBM device are subject to mild non-Markovian dissipation with spatial correlations.

quant-ph↗

arXiv2019

Rethinking the Usage of Batch Normalization and Dropout in the Training of Deep Neural Networks

In this work, we propose a novel technique to boost training efficiency of a neural network. Our work is based on an excellent idea that whitening the inputs of neural networks can achieve a fast convergence speed. Given the well-known fact that independent components must be whitened, we introduce a novel Independent-Component (IC) layer before each weight layer, whose inputs would be made more independent. However, determining independent components is a computationally intensive task. To overcome this challenge, we propose to implement an IC layer by combining two popular techniques, Batch Normalization and Dropout, in a new manner that we can rigorously prove that Dropout can quadratically reduce the mutual information and linearly reduce the correlation between any pair of neurons with respect to the dropout layer parameter $p$. As demonstrated experimentally, the IC layer consistently outperforms the baseline approaches with more stable training process, faster convergence speed and better convergence limit on CIFAR10/100 and ILSVRC2012 datasets. The implementation of our IC layer makes us rethink the common practices in the design of neural networks. For example, we should not place Batch Normalization before ReLU since the non-negative responses of ReLU will make the weight layer updated in a suboptimal way, and we can achieve better performance by combining Batch Normalization and Dropout together as an IC layer.

cs.LG↗

arXiv2018

Heat transfer statistics in mixed quantum-classical systems

The modelling of quantum heat transfer processes at the nanoscale is crucial for the development of energy harvesting and molecular electronics devices. Herein, we adopt a mixed quantum-classical description of a device, in which the open subsystem of interest is treated quantum mechanically and the surrounding heat baths are treated in a classical-like fashion. By introducing such a mixed quantum-classical description of the composite system, one is able to study the heat transfer between the subsystem and bath from a closed system point of view, thereby avoiding simplifying assumptions related to the bath time scale and subsystem-bath coupling strength. In particular, we adopt the full counting statistics approach to derive a general expression for the moment generating function of heat in systems whose dynamics are described by the quantum-classical Liouville equation (QCLE). From this expression, one can deduce expressions for the dynamics of the average heat and heat current, which may be evaluated using numerical simulations. Due to the approximate nature of the QCLE, we also find that the steady state fluctuation symmetry holds up to order $\hbar$ for systems whose subsystembath couplings and baths go beyond bilinear and harmonic, respectively. To demonstrate the approach, we consider the nonequilibrium spin boson model and simulate its time-dependent average heat and heat current under various conditions.

physics.chem-ph↗

arXiv2018

Optimizing Quantum Walk Search on a Reduced Uniform Complete Multi-Partite Graph

In a recent work by Novo et al. (Sci. Rep. 5, 13304, 2015), the invariant subspace method was applied to the study of continuous-time quantum walk (CTQW). The method helps to reduce a graph into a simpler version that allows more transparent analyses of the quantum walk model. In this work, we adopt the aforementioned method to investigate the optimality of a quantum walk search of a marked element on a complete multi-partite graph. We formulate the eigenbasis that would facilitate the transport between the two lowest energy eigenstates and demonstrate how to set the appropriate coupling factor to preserve the optimality.

quant-ph↗

arXiv2017

Frequency-dependent current noise in quantum heat transfer with full counting statistics

To investigate frequency-dependent current noise (FDCN) in open quantum systems at steady states, we present a theory which combines Markovian quantum master equations with a finite time full counting statistics. Our formulation of the FDCN generalizes previous zero-frequency expressions and can be viewed as an application of MacDonald's formula for electron transport to heat transfer. As a demonstration, we consider the paradigmatic example of quantum heat transfer in the context of a non-equilibrium spin-boson model. We adopt a recently developed polaron-transformed Redfield equation which allows us to accurately investigate heat transfer with arbitrary system-reservoir coupling strength, arbitrary values of spin bias as well as temperature differences. We observe maximal values of FDCN in moderate coupling regimes, similar to the zero-frequency cases. We find the FDCN with varying coupling strengths or bias displays a universal Lorentzian-shape scaling form in the weak coupling regime, and a white noise spectrum emerges with zero bias in the strong coupling regime due to a distinctive spin dynamics. We also find the bias can suppress the FDCN in the strong coupling regime, in contrast to its zero-frequency counterpart which is insensitive to bias changes. Furthermore, we utilize the Saito-Utsumi relation as a benchmark to validate our theory and study the impact of temperature differences at finite frequencies. Together, our results provide detailed dissections of the finite time fluctuation of heat current in open quantum systems.

cond-mat.stat-mech↗

arXiv2017

Efficiency at maximum power of a quantum Carnot engine with temperature tunable baths

We investigate the efficiency at maximum power (EMP) of irreversible quantum Carnot engines that perform finite-time cycles between two temperature tunable baths. The temperature form we adopt can be experimentally realized in squeezed baths in the high temperature limit, which makes our proposal of practical relevance. Focusing on low dissipation engines, we first generalize the pervious upper as well as lower bounds for the EMP to temperature tunable cases in which they are solely determined by a generalized Carnot limit. As an illustrative example, we then consider a minimal heat engine model with a two-level spin as the working medium. It mimics a low dissipation engine as confirmed by finite time thermodynamic optimization results. The so-obtained EMP, being constrained by the generalized bounds, is well described by a generalized Curzon- Ahlborn efficiency as consequences of a left/right symmetry for a rate constant and low dissipations. Intriguing features of this minimal heat engine under optimal power output are also demonstrated.

quant-ph↗

arXiv2017

Ultrasensitive and highly accurate long-range surface plasmon resonance biosensors based on two-dimensional transition metal dichalcogenides

Two-dimensional transition metal dichalcogenides (TMDCs), as promising alternative plasmonics supporting materials to graphene, exhibit potential applications in sensing. Here, we propose an ultrasensitive, accurate long-range surface plasmon resonance (LRSPR) imaging biosensor with two-dimensional TMDC layers, which shows higher detection accuracy than that of conventional SPR biosensor. It is found that the imaging sensitivity of the proposed LRSPR biosensor can be enhanced by the integration of TMDC layers, which is different from the previous graphene-based LRSPR or SPR imaging sensor, whose imaging sensitivity usually decreases with the number of graphene layers. The sensitivity enhancement or degradation effect for the proposed chalcogenide-cytop-gold-TMDCs based biosensor depends on the thickness of gold thin film and cytop layer. Imaging sensitivity of more than 4000 $\text{RIU}^{-1}$ can be obtained with a high detection accuracy of more than 120 $\text{deg}^{-1}$. We expect that the proposed TMDCs mediated LRSPR imaging sensor could provide potential applications in chemical sensing and biosensing for a highly sensitive and accurate simultaneous detection of multiple biomolecular interactions.

physics.app-ph↗

arXiv2017

A Unified Stochastic Formulation of Dissipative Quantum Dynamics. I. Generalized Hierarchical Equations

We extend a standard stochastic theory to study open quantum systems coupled to generic quantum environments including the three fundamental classes of noninteracting particles: bosons, fermions and spins. In this unified stochastic approach, the generalized stochastic Liouville equation (SLE) formally captures the exact quantum dissipations when noise variables with appropriate statistics for different bath models are applied. Anharmonic effects of a non-Gaussian bath are precisely encoded in the bath multi-time correlation functions that noise variables have to satisfy. Staring from the SLE, we devise a family of generalized hierarchical equations by averaging out the noise variables and expand bath multi-time correlation functions in a complete basis of orthonormal functions. The general hiearchical equations constitute systems of linear equations that provide numerically exact simulations of quantum dynamics. For bosonic bath models, our general hierarchical equation of motion reduces exactly to an extended version of hierarchical equation of motion which allows efficient simulation for arbitrary spectral densities and temperature regimes. Similar efficiency and exibility can be achieved for the fermionic bath models within our formalism. The spin bath models can be simulated with two complementary approaches in the presetn formalism. (I) They can be viewed as an example of non-Gaussian bath models and be directly handled with the general hierarchical equation approach given their multi-time correlation functions. (II) Alterantively, each bath spin can be first mapped onto a pair of fermions and be treated as fermionic environments within the present formalism.

physics.chem-ph↗

arXiv2017

A Unified Stochastic Formulation of Dissipative Quantum Dynamics. II. Beyond Linear Response of Spin Baths

We use the "generalized hierarchical equation of motion" proposed in Paper I to study decoherence in a system coupled to a spin bath. The present methodology allows a systematic incorporation of higher order anharmonic effects of the bath in dynamical calculations. We investigate the leading order corrections to the linear response approximations for spin bath models. Two types of spin-based environments are considered: (1) a bath of spins discretized from a continuous spectral density and (2) a bath of physical spins such as nuclear or electron spins. The main difference resides with how the bath frequency and the system-bath coupling parameters are chosen to represent an environment. When discretized from a continuous spectral density, the system-bath coupling typically scales as $\sim 1/\sqrt{N_B}$ where $N_B$ is the number of bath spins. This scaling suppresses the non-Gaussian characteristics of the spin bath and justify the linear response approximations in the thermodynamic limit. For the physical spin bath models, system-bath couplings are directly deduced from spin-spin interactions with no reason to obey the $1/\sqrt{N_B}$ scaling. It is not always possible to justify the linear response approximations. Furthermore, if the spin-spin Hamiltonian and/or the bath parameters are highly symmetrical, these additional constraints generate non-Markovian and persistent dynamics that is beyond the linear response treatments.

physics.chem-ph↗

arXiv2013

Analysis of geometric phase effects in the quantum-classical Liouville formalism

We analyze two approaches to the quantum-classical Liouville (QCL) formalism that differ in the order of two operations: Wigner transformation and projection onto adiabatic electronic states. The analysis is carried out on a two-dimensional linear vibronic model where geometric phase (GP) effects arising from a conical intersection profoundly affect nuclear dynamics. We find that the Wigner-then-Adiabatic (WA) QCL approach captures GP effects, whereas the Adiabatic-then-Wigner (AW) QCL approach does not. Moreover, the Wigner transform in AW-QCL leads to an ill-defined Fourier transform of double-valued functions. The double-valued character of these functions stems from the nontrivial GP of adiabatic electronic states in the presence of a conical intersection. In contrast, WA-QCL avoids this issue by starting with the Wigner transform of single-valued quantities of the full problem. Since the WA-QCL approach uses solely the adiabatic potentials and non-adiabatic derivative couplings as an input, our results indicate that WA-QCL can capture GP effects in general two-state crossing problems using first-principles electronic structure calculations without prior diabatization or introduction of explicit phase factors.

quant-ph↗

arXiv2013

Analysis of the Forward-Backward Trajectory Solution for the Mixed Quantum-Classical Liouville Equation

Mixed quantum-classical methods provide powerful algorithms for the simulation of quantum processes in large and complex systems. The forward-backward trajectory solution of the mixed quantum-classical Liouville equation in the mapping basis [J. Chem. Phys. 137, 22A507 (2012)] is one such scheme. It simulates the dynamics via the propagation of forward and backward trajectories of quantum coherent state variables, and the propagation of bath trajectories on a mean-field potential determined jointly by the forward and backward trajectories. An analysis of the properties of this solution, numerical tests of its validity and an investigation of its utility for the study of nonadiabtic quantum processes are given. In addition, we present an extension of this approximate solution that allows one to systematically improve the results. This extension, termed the jump forward-backward trajectory solution, is analyzed and tested in detail and its various implementations are discussed.

physics.chem-ph↗

arXiv2012

Nonadiabatic Dynamics in Open Quantum-Classical Systems: Forward-Backward Trajectory Solution

A new approximate solution to the quantum-classical Liouville equation is derived starting from the formal solution of this equation in forward-backward form. The time evolution of a mixed quantum-classical system described by this equation is obtained in a coherent state basis using the mapping representation, which expresses $N$ quantum degrees of freedom in a 2N-dimensional phase space. The solution yields a simple non-Hamiltonian dynamics in which a set of $N$ coherent state coordinates evolve in forward and backward trajectories while the bath coordinates evolve under the influence of the mean potential that depends on these forward and backward trajectories. It is shown that the solution satisfies the differential form of the quantum-classical Liouville equation exactly. Relations to other mixed quantum-classical and semi-classical schemes are discussed.

physics.chem-ph↗

arXiv2011

Theory of electronic properties and quantum spin blockade in a gated linear triple quantum dot with one electron spin each

We present a theory of electronic properties and the spin blockade phenomena in a gated linear triple quantum dot. Quadruple points where four different charge configurations are on resonance, particularly involving (1,1,1) configuration, are considered. In the symmetric case, the central dot is biased to higher energy and a single electron tunnels through the device when (1,1,1) configuration is resonant with (1,0,1),(2,0,1),(1,0,2) configurations. The electronic properties of a triple quantum dot are described by a Hubbard model containing two orbitals in the two unbiased dots and a single orbital in the biased dot. The transport through the triple quantum dot molecule involves both singly and doubly occupied configurations and necessitates the description of the (1,1,1) configuration beyond the Heisenberg model. Exact eigenstates of the triple quantum dot molecule with up to three electrons are used to compute current assuming weak coupling to the leads and non-equilibrium occupation of quantum molecule states obtained from the rate equation. The intra-molecular relaxation processes due to acoustic phonons and cotunneling with the leads are included, and are shown to play a crucial role in the spin blockade effect. We find a quantum interference-based spin blockade phenomenon at low source-drain bias and a distinct spin blockade due to a trap state at higher bias. We also show that, for an asymmetric quadruple point with (0,1,1),(1,1,1,),(0,2,1),(0,1,2) configurations on resonance, the spin blockade is analogous to the spin blockade in a double quantum dot.

cond-mat.mes-hall↗

arXiv2011

Herzberg Circuit and Berry's Phase in Chirality-based Coded Qubit in a Triangular Triple Quantum Dot

We present a theoretical proposal for the Herzberg circuit and controlled accumulation of Berry's phase in a chirality-based coded qubit in a triangular triple quantum dot molecule with one electron spin each. The qubit is encoded in the two degenerate states of a three spin complex with total spin $S=1/2$. Using a Hubbard and Heisenberg model the Herzberg circuit encircling the degeneracy point is realized by adiabatically tuning the successive on-site energies of quantum dots and tunnel couplings across a pair of neighbouring dots. It is explicitly shown that encircling the degeneracy point leads to the accumulation of the geometrical Berrys phase. We show that only triangular but not linear quantum dot molecule allows for the generation of Berry's phase and we discuss a protocol to detect this geometrical phase.

cond-mat.mes-hall↗

arXiv2010

Valence holes as Luttinger spinor based qubits in quantum dots

We present a theory of valence holes as Luttinger spinor based qubits in p-doped self-assembled quantum dots within the 4-band $k\cdot p$ formalism. The two qubit levels are identified with the two chiralities of the doubly degenerate ground state. We show that single qubit operations can be implemented with static magnetic field applied along the $z$ and $x$ directions, acting analogously to the $\hatσ_z$ and $\hatσ_x$ operators in the qubit subspace respectively. The coupling of two dots and hence the double qubit operations are shown to be sensitive to the orientation of the two quantum dots. For vertical qubit arrays, there exists an optimal qubit separation suitable for the voltage control of qubit-qubit interactions.

cond-mat.str-el↗

arXiv2010

Quantum circuits based on coded qubits encoded in chirality of electron spin complexes in triple quantum dots

We present a theory of quantum circuits based on logical qubits encoded in chirality of electron spin complexes in lateral gated semiconductor triple quantum dot molecules with one electron spin in each dot. Using microscopic Hamiltonian we show how to initialize, coherently control and measure the quantum state of a chirality based coded qubit using static in-plane magnetic field and voltage tuning of individual dots. The microscopic model of two interacting coded qubits is established and mapped to an Ising Hamiltonian, resulting in conditional two-qubit phase gate.

cond-mat.str-el↗