arXiv subjects

Bo Peng

Publications and source records attributed to Bo Peng.

At least 307 records · Page 17Linked to original sources

arXiv2022

UWC: Unit-wise Calibration Towards Rapid Network Compression

This paper introduces a post-training quantization~(PTQ) method achieving highly efficient Convolutional Neural Network~ (CNN) quantization with high performance. Previous PTQ methods usually reduce compression error via performing layer-by-layer parameters calibration. However, with lower representational ability of extremely compressed parameters (e.g., the bit-width goes less than 4), it is hard to eliminate all the layer-wise errors. This work addresses this issue via proposing a unit-wise feature reconstruction algorithm based on an observation of second order Taylor series expansion of the unit-wise error. It indicates that leveraging the interaction between adjacent layers' parameters could compensate layer-wise errors better. In this paper, we define several adjacent layers as a Basic-Unit, and present a unit-wise post-training algorithm which can minimize quantization error. This method achieves near-original accuracy on ImageNet and COCO when quantizing FP32 models to INT4 and INT3.

cs.CV↗

arXiv2021

Phonons as a platform for non-Abelian braiding and its manifestation in layered silicates

Topological phases of matter have revolutionised the fundamental understanding of band theory and hold great promise for next-generation technologies such as low-power electronics or quantum computers. Single-gap topologies have been extensively explored, and a large number of materials have been theoretically proposed and experimentally observed. These ideas have recently been extended to multi-gap topologies with band nodes that carry non-Abelian charges, characterised by invariants that arise by the momentum space braiding of such nodes. However, the constraints placed by the Fermi-Dirac distribution to electronic systems have so far prevented the experimental observation of multi-gap topologies in real materials. Here, we show that multi-gap topologies and the accompanying phase transitions driven by braiding processes can be readily observed in the bosonic phonon spectra of known monolayer silicates. The associated braiding process can be controlled by means of an electric field and epitaxial strain, and involves, for the first time, more than three bands. Finally, we propose that the band inversion processes at the $Γ$ point can be tracked by following the evolution of the Raman spectrum, providing a clear signature for the experimental verification of the band inversion accompanied by the braiding process.

cond-mat.mes-hall↗

arXiv2021

Quantum time dynamics of 1D-Heisenberg models employing the Yang-Baxter equation for circuit compression

Quantum time dynamics (QTD) is considered a promising problem for quantum supremacy on near-term quantum computers. However, QTD quantum circuits grow with increasing time simulations. This study focuses on simulating the time dynamics of 1-D integrable spin chains with nearest neighbor interactions. We show how the quantum Yang-Baxter equation can be exploited to compress and produce a shallow quantum circuit. With this compression scheme, the depth of the quantum circuit becomes independent of step size and only depends on the number of spins. We show that the compressed circuit scales quadratically with system size, which allows for the simulations of time dynamics of very large 1-D spin chains. We derive the compressed circuit representations for different special cases of the Heisenberg Hamiltonian. We compare and demonstrate the effectiveness of this approach by performing simulations on quantum computers.

quant-ph↗

arXiv2021

Multi-View Stereo with Transformer

This paper proposes a network, referred to as MVSTR, for Multi-View Stereo (MVS). It is built upon Transformer and is capable of extracting dense features with global context and 3D consistency, which are crucial to achieving reliable matching for MVS. Specifically, to tackle the problem of the limited receptive field of existing CNN-based MVS methods, a global-context Transformer module is first proposed to explore intra-view global context. In addition, to further enable dense features to be 3D-consistent, a 3D-geometry Transformer module is built with a well-designed cross-view attention mechanism to facilitate inter-view information interaction. Experimental results show that the proposed MVSTR achieves the best overall performance on the DTU dataset and strong generalization on the Tanks & Temples benchmark dataset.

cs.CV↗

arXiv2021

EQC : Ensembled Quantum Computing for Variational Quantum Algorithms

Variational quantum algorithm (VQA), which is comprised of a classical optimizer and a parameterized quantum circuit, emerges as one of the most promising approaches for harvesting the power of quantum computers in the noisy intermediate scale quantum (NISQ) era. However, the deployment of VQAs on contemporary NISQ devices often faces considerable system and time-dependant noise and prohibitively slow training speeds. On the other hand, the expensive supporting resources and infrastructure make quantum computers extremely keen on high utilization. In this paper, we propose a virtualized way of building up a quantum backend for variational quantum algorithms: rather than relying on a single physical device which tends to introduce temporal-dependant device-specific noise with worsening performance as time-since-calibration grows, we propose to constitute a quantum ensemble, which dynamically distributes quantum tasks asynchronously across a set of physical devices, and adjusting the ensemble configuration with respect to machine status. In addition to reduced machine-dependant noise, the ensemble can provide significant speedups for VQA training. With this idea, we build a novel VQA training framework called EQC that comprises: (i) a system architecture for asynchronous parallel VQA cooperative training; (ii) an analytic model for assessing the quality of the returned VQA gradient over a particular device concerning its architecture, transpilation, and runtime conditions; (iii) a weighting mechanism to adjust the quantum ensemble's computational contribution according to the systems' current performance. Evaluations comprising 500K circuit evaluations across 10 IBMQ devices using a VQE and a QAOA applications demonstrate that EQC can attain error rates close to the most performant device of the ensemble, while boosting the training speed by 10.5x on average (up to 86x and at least 5.2x).

quant-ph↗

arXiv2021

Phonon-assisted electronic states modulation of few-layer PdSe2 at terahertz frequencies

Information technology demands high-speed optoelectronic devices, but going beyond the one terahertz (THz) barrier is challenging due to the difficulties associated with generating, detecting, and processing high-frequency signals. Here, we show that femtosecond-laser-driven phonons can be utilized to coherently manipulate the excitonic properties of semiconductors at THz frequencies. The precise control of the pump and subsequent time-delayed broadband probe pulses enables the simultaneous generation and detection processes of both periodic lattice vibrations and their couplings with electronic states. Combining ultralow frequency Raman spectroscopy with first-principles calculations, we identify the unique phonon mode-selective and probe-energy dependent features of electron-phonon interactions in layered PdSe2. Two distinctive types of coherent phonon excitations could couple preferentially to different types of electronic excitations: the intralayer (4.3 THz) mode to carriers and the interlayer (0.35 THz) mode to excitons. This work provides new insights to understand the excited-state phonon interactions of 2D materials, and to achieve future applications of optoelectronic devices operating at THz frequencies.

cond-mat.mes-hall↗

arXiv2021

Degenerate topological line surface phonons in quasi-1D double helix crystal SnIP

Degenerate points/lines in the bulk band structures of crystals have become a staple of the growing number of topological materials. The bulk-boundary correspondence provides a relation between bulk topology and surface states. While line degeneracies of bulk excitations have been extensively characterized, line degeneracies of surface states are not well understood. We show that SnIP, a quasi-one-dimensional van der Waals material with a double helix crystal structure, exhibits topological nodal rings/lines in both the bulk phonon modes and their corresponding surface states. Using a combination of first-principles calculations, symmetry-based indicator theories and Zak phase analysis, we find that two neighbouring bulk nodal rings form doubly degenerate lines in their drumhead-like surface states, which are protected by the combination of time-reversal and glide mirror symmetries $\mathcal{T}\bar{M}_y$. Our results indicate that surface degeneracies can be generically protected by symmetries such as $\mathcal{T}\bar{M}_y$, and phonons provide an ideal platform to explore such degeneracies.

cond-mat.mtrl-sci↗

arXiv2021

Multi-gap topology and non-Abelian braiding of phonons from first principles

Non-Abelian states of matter, in which the final state depends on the order of the interchanges of two quasiparticles, can encode information immune from environmental noise with the potential to provide a robust platform for topological quantum computation. We demonstrate that phonons can carry non-Abelian frame charges at the band crossing points of their frequency spectrum, and that external stimuli can drive their braiding. We present a general framework to understand the topological configurations of phonons from first principles calculations using a topological invariant called Euler class, and provide a complete analysis of phonon braiding by combining different topological configurations. Taking a well-known dielectric material, Al$_2$O$_3$, as a representative example, we demonstrate that electrostatic doping gives rise to phonon band inversions that can induce redistribution of the frame charges, leading to non-Abelian braiding of phonons. Our work provides a new quasiparticle platform for realizable non-Abelian braiding in reciprocal space, and expands the toolset for studying braiding processes.

cond-mat.mes-hall↗

arXiv2021

Novel View Synthesis from a Single Image via Unsupervised learning

View synthesis aims to generate novel views from one or more given source views. Although existing methods have achieved promising performance, they usually require paired views of different poses to learn a pixel transformation. This paper proposes an unsupervised network to learn such a pixel transformation from a single source viewpoint. In particular, the network consists of a token transformation module (TTM) that facilities the transformation of the features extracted from a source viewpoint image into an intrinsic representation with respect to a pre-defined reference pose and a view generation module (VGM) that synthesizes an arbitrary view from the representation. The learned transformation allows us to synthesize a novel view from any single source viewpoint image of unknown pose. Experiments on the widely used view synthesis datasets have demonstrated that the proposed network is able to produce comparable results to the state-of-the-art methods despite the fact that learning is unsupervised and only a single source viewpoint image is required for generating a novel view. The code will be available soon.

cs.CV↗

arXiv2021

Statistical tests of young radio pulsars with/without supernova remnants: implying two origins of neutron stars

The properties of the young pulsars and their relations to the supernova remnants (SNRs) have been the interesting topics. At present, 383 SNRs in the Milky Way galaxy have been published, which are associated with 64 radio pulsars and 46 pulsars with high energy emissions. However, we noticed that 630 young radio pulsars with spin periods of less than half a second have been not yet observed the SNRs surrounding or nearby them, which arises a question of that could the two types of young radio pulsars with/without SNRs hold distinctive characteristics? Here, we employ the statistical tests on the two groups of young radio pulsars with (52) and without (630) SNRs to reveal if they share different origins. Kolmogorov-Smirnov (K-S) and Mann-Whitney-Wilcoxon (M-W-W) tests indicate that the two samples have the different distributions with parameters of spin period ($P$), derivative of spin period ($\dot P$), surface magnetic field strength ($B$), and energy loss rate ($\dot E$). Meanwhile, the cumulative number ratio between the pulsars with and without SNRs at the different spindown ages decreases significantly after $\rm10-20\,Kyr$. So we propose that the existence of the two types of supernovae (SNe), corresponding to their SNR lifetimes, which can be roughly ascribed to the low-energy and high-energy SNe. Furthermore, the low-energy SNe may be formed from the $\rm8-12\,M_{\odot}$ progenitor, e.g., possibly experiencing the electron capture, while the main sequence stars of $\rm12-25\,M_{\odot}$ may produce the high-energy SNe probably by the iron core collapse.

astro-ph.HE↗

arXiv2021

State preparation and evolution in quantum computing: a perspective from Hamiltonian moments

Quantum algorithms on the noisy intermediate-scale quantum (NISQ) devices are expected to simulate quantum systems that are classically intractable to demonstrate quantum advantages. However, the non-negligible gate error on the NISQ devices impedes the conventional quantum algorithms to be implemented. Practical strategies usually exploit hybrid quantum classical algorithms to demonstrate potentially useful applications of quantum computing in the NISQ era. Among the numerous hybrid algorithms, recent efforts highlight the development of quantum algorithms based upon quantum computed Hamiltonian moments, $\langle ϕ| \hat{\mathcal{H}}^n | ϕ\rangle$ ($n=1,2,\cdots$), with respect to quantum state $|ϕ\rangle$. In this tutorial, we will give a brief review of these quantum algorithms with focuses on the typical ways of computing Hamiltonian moments using quantum hardware and improving the accuracy of the estimated state energies based on the quantum computed moments. Furthermore, we will present a tutorial to show how we can measure and compute the Hamiltonian moments of a four-site Heisenberg model, and compute the energy and magnetization of the model utilizing the imaginary time evolution in the real IBM-Q NISQ hardware environment. Along this line, we will further discuss some practical issues associated with these algorithms. We will conclude this tutorial review by overviewing some possible developments and applications in this direction in the near future.

quant-ph↗

arXiv2021

Induced transparency: interference or polarization?

The polarization of optical fields is a crucial degree of freedom in the all-optical analogue of electromagnetically induced transparency (EIT). However, the physical origins of EIT and polarization induced phenomena have not been well distinguished, which can lead to confusion in associated applications such as slow light and optical/quantum storage. Here we study the polarization effects in various optical EIT systems. We find that a polarization mismatch between whispering gallery modes in two indirectly coupled resonators can induce a narrow transparency window in the transmission spectrum resembling the EIT lineshape. However, such polarization induced transparency (PIT) is distinct from EIT: it originates from strong polarization rotation effects and shows unidirectional feature. The coexistence of PIT and EIT provides new routes for the manipulation of light flow in optical resonator systems.

physics.optics↗

arXiv2021

Large Monochromatic Components of Small Diameter

Gyárfás conjectured in 2011 that every $r$-edge-colored $K_n$ contains a monochromatic component of bounded ("perhaps three") diameter on at least $n/(r-1)$ vertices. Letzter proved this conjecture with diameter four. In this note we improve the result in the case of $r=3$: We show that in every $3$-edge-coloring of $K_n$ either there is a monochromatic component of diameter at most three on at least $n/2$ vertices or every color class is spanning and has diameter at most four.

math.CO↗

arXiv2021

VQE Method: A Short Survey and Recent Developments

The variational quantum eigensolver (VQE) is a method that uses a hybrid quantum-classical computational approach to find eigenvalues and eigenvalues of a Hamiltonian. VQE has been proposed as an alternative to fully quantum algorithms such as quantum phase estimation because fully quantum algorithms require quantum hardware that will not be accessible in the near future. VQE has been successfully applied to solve the electronic Schrödinger equation for a variety of small molecules. However, the scalability of this method is limited by two factors: the complexity of the quantum circuits and the complexity of the classical optimization problem. Both of these factors are affected by choice of the variational ansatz used to represent the trial wave function. Hence, the construction of efficacious ansatz is an active area of research. Put another way, modern quantum computers are not capable of executing deep quantum circuits produced by using currently available ansatze for problems that map onto more than several qubits. In this review, we present recent developments in the field of designing effective ansatzes that fall into two categories -- chemistry inspired and hardware efficient -- that produce quantum circuits that are easier to run on modern hardware. We discuss the shortfalls of ansatzes originally formulated for VQE simulations, how they are addressed in more sophisticated methods, and the potential ways for further improvements.

quant-ph↗

arXiv2021

Coupled cluster Green's function-Past, Present, and Future

Coupled cluster Green's function (CCGF) approach has drawn much attention in recent years for targeting the molecular and material electronic structure problems from a many-body perspective in a systematically improvable way. Here, we will present a brief review of the history of how the Green's function method evolved with the wavefunction, early and recent development of CCGF theory, and more recently scalable CCGF software development. We will highlight some of the recent applications of CCGF approach and propose some potential applications that would emerge in the near future.

physics.chem-ph↗

arXiv2021

SOGAN: 3D-Aware Shadow and Occlusion Robust GAN for Makeup Transfer

In recent years, virtual makeup applications have become more and more popular. However, it is still challenging to propose a robust makeup transfer method in the real-world environment. Current makeup transfer methods mostly work well on good-conditioned clean makeup images, but transferring makeup that exhibits shadow and occlusion is not satisfying. To alleviate it, we propose a novel makeup transfer method, called 3D-Aware Shadow and Occlusion Robust GAN (SOGAN). Given the source and the reference faces, we first fit a 3D face model and then disentangle the faces into shape and texture. In the texture branch, we map the texture to the UV space and design a UV texture generator to transfer the makeup. Since human faces are symmetrical in the UV space, we can conveniently remove the undesired shadow and occlusion from the reference image by carefully designing a Flip Attention Module (FAM). After obtaining cleaner makeup features from the reference image, a Makeup Transfer Module (MTM) is introduced to perform accurate makeup transfer. The qualitative and quantitative experiments demonstrate that our SOGAN not only achieves superior results in shadow and occlusion situations but also performs well in large pose and expression variations.

cs.CV↗

arXiv2021

Variational quantum solver employing the PDS energy functional

Recently a new class of quantum algorithms that are based on the quantum computation of the connected moment expansion has been reported to find the ground and excited state energies. In particular, the Peeters-Devreese-Soldatov (PDS) formulation is found variational and bearing the potential for further combining with the existing variational quantum infrastructure. Here we find that the PDS formulation can be considered as a new energy functional of which the PDS energy gradient can be employed in a conventional variational quantum solver. In comparison with the usual variational quantum eigensolver (VQE) and the original static PDS approach, this new variational quantum solver offers an effective approach to navigate the dynamics to be free from getting trapped in the local minima that refer to different states, and achieve high accuracy at finding the ground state and its energy through the rotation of the trial wave function of modest quality, thus improves the accuracy and efficiency of the quantum simulation. We demonstrate the performance of the proposed variational quantum solver for toy models, H$_2$ molecule, and strongly correlated planar H$_4$ system in some challenging situations. In all the case studies, the proposed variational quantum approach outperforms the usual VQE and static PDS calculations even at the lowest order. We also discuss the limitations of the proposed approach and its preliminary execution for model Hamiltonian on the NISQ device.

quant-ph↗

arXiv2021

DFGC 2021: A DeepFake Game Competition

This paper presents a summary of the DFGC 2021 competition. DeepFake technology is developing fast, and realistic face-swaps are increasingly deceiving and hard to detect. At the same time, DeepFake detection methods are also improving. There is a two-party game between DeepFake creators and detectors. This competition provides a common platform for benchmarking the adversarial game between current state-of-the-art DeepFake creation and detection methods. In this paper, we present the organization, results and top solutions of this competition and also share our insights obtained during this event. We also release the DFGC-21 testing dataset collected from our participants to further benefit the research community.

cs.CV↗