arXiv Science⌕ Search

arXiv subjects

Hui Chen

Publications and source records attributed to Hui Chen.

At least 343 records · Page 19Linked to original sources

Atomically-Precise, Custom-Design Origami Graphene Nanostructures

The construction of atomically-precise carbon nanostructures holds promise for developing novel materials for scientific study and nanotechnology applications. Here we show that graphene origami is an efficient way to convert graphene into atomically-precise, complex, and novel nanostructures. By scanning-tunneling-microscope manipulation at low temperature, we repeatedly fold and unfold graphene nanoislands (GNIs) along arbitrarily chosen direction. A bilayer graphene stack featuring a tunable twist angle and a tubular edge connection between the layers are formed. Folding single-crystal GNIs creates tubular edges with specified chirality and one-dimensional electronic features similar to those of carbon nanotubes, while folding bi-crystal GNIs creates well-defined intramolecular junctions. Both origami structural models and electronic band structures were computed to complement analysis of the experimental results. The present atomically-precise graphene origami provides a platform for constructing novel carbon nanostructures with engineered quantum properties and ultimately quantum machines.

cond-mat.mtrl-sci↗

Microtwist elasticity: A continuum approach to zero modes and topological polarization in Kagome lattices

The topologically polarized isostatic lattices discovered by Kane and Lubensky (2014, Nat. Phys. 10, 39-45) challenged the standard effective medium theories used in the modeling of many truss-based materials and metamaterials. As a matter of fact, these exhibit Parity (P) asymmetric distributions of zero modes that induce a P-asymmetric elastic behavior, both of which cannot be reproduced within Cauchy elasticity. Here, we propose a new effective medium theory baptized "microtwist elasticity" capable of rendering polarization effects on a macroscopic scale. The theory is valid for trusses on the brink of a polarized-unpolarized phase transition in which case they necessarily exhibit more periodic zero modes than they have dimensions. By mapping each periodic zero mode to a macroscopic degree of freedom, the microtwist theory ends up being a kinematically enriched theory. Microtwist elasticity is constructed thanks to leading order two-scale asymptotics and its constitutive and balance equations are derived for a fairly generic isostatic truss: the Kagome lattice. Various numerical and analytical calculations, of the shape and distribution of zero modes, of dispersion diagrams and of polarization effects, systematically show the quality of the proposed effective medium theory. Most notably, the theory is capable of producing a continuum version of Kane and Lubensky's topological polarization vector.

physics.class-ph↗

IMRAM: Iterative Matching with Recurrent Attention Memory for Cross-Modal Image-Text Retrieval

Enabling bi-directional retrieval of images and texts is important for understanding the correspondence between vision and language. Existing methods leverage the attention mechanism to explore such correspondence in a fine-grained manner. However, most of them consider all semantics equally and thus align them uniformly, regardless of their diverse complexities. In fact, semantics are diverse (i.e. involving different kinds of semantic concepts), and humans usually follow a latent structure to combine them into understandable languages. It may be difficult to optimally capture such sophisticated correspondences in existing methods. In this paper, to address such a deficiency, we propose an Iterative Matching with Recurrent Attention Memory (IMRAM) method, in which correspondences between images and texts are captured with multiple steps of alignments. Specifically, we introduce an iterative matching scheme to explore such fine-grained correspondence progressively. A memory distillation unit is used to refine alignment knowledge from early steps to later ones. Experiment results on three benchmark datasets, i.e. Flickr8K, Flickr30K, and MS COCO, show that our IMRAM achieves state-of-the-art performance, well demonstrating its effectiveness. Experiments on a practical business advertisement dataset, named \Ads{}, further validates the applicability of our method in practical scenarios.

cs.CV↗

Realization of ppm level pressure stability for primary thermometry using a primary piston gauge

To achieve an uncertainty of 0.25 mK in single-pressure refractive-index gas thermometry (SPRIGT), the relative pressure variation of He-4 gas in the range 30 kPa to 90 kPa, should not exceed 4 ppm (k=1). To this end, a novel pressure control system has been developed. It consists of two main parts: a piston gauge to control the pressure, and a home-made gas compensation system to supplement the micro-leak of the piston gauge. In addition, to maintain the piston at constant height, a servo loop is used that automatically determines in real time the amount of extra gas required. At room temperature, the standard deviations of the stabilized pressure are 3.0 mPa at 30 kPa, 4.5 mPa at 60 kPa and 2 mPa at 90 kPa. For the temperature region 5 K-25 K used for SPRIGT in the present work, the relative pressure stability is better than 0.16 ppm i.e. 25 times better than required. Moreover, the same pressure stabilization system is readily transposable to other primary gas thermometers.

physics.ins-det↗

Active suppression of temperature oscillation from a pulse-tube cryocooler in a cryogen-free cryostat: Part 2. Experimental realization

A cryogen-free cryostat cooled by a closed cycle cryocooler is compact, can provide uninterrupted long-term operation (up to ten thousand hours) and is suited to temperatures from 3 K to 300 K. Its intrinsic temperature oscillation, however, limits its application in experiments requiring high thermal stability at low temperature (below 77 K). Passive suppression methods are effective but all suffer from drawbacks. We describe a novel, active suppression scheme more efficient than traditional proportional-integral (PI) control. The experimental results show that it can reduce the standard deviation of the temperature oscillation by a further 30% compared with PI feedback. To the best of our knowledge, this is the first time such active suppression of temperature oscillations has been implemented with the cryogen-free cryostat. The results also show, however, that an unwanted lower frequency thermal noise will be generated, which appears to be the limit of the method. Nevertheless, the approach could be used to improve the temperature stability in all cryogen-free cryostats.

physics.ins-det↗

Active suppression of temperature oscillation from a pulse-tube cryocooler in a cryogen-free cryostat: Part 1. Simulation modeling from thermal response characteristics

A cryogen-free cryostat cooled using a 4 K commercial GM or pulse tube cryocooler (PTC) displays temperature oscillations caused by the intrinsic working principle of the regenerative cryocooler. To dampen such oscillations usually requires either a large heat capacity or a large thermal resistance. To understand this phenomenon better and suppress it more effectively, both the step response characteristic and the intrinsic oscillation characteristic of cryostat have been used to obtain the complete transfer functions of a simulation model. The latter is used to test and optimize traditional PID feedback control. The results showed this approach has almost no effect on the temperature oscillation amplitude. Based on this simulation model, a novel active method was proposed and tested numerically. Simulation results predict the method should suppress the amplitude of the original temperature oscillation by a factor of two.

physics.ins-det↗

Nearly quantized conductance plateau of vortex zero mode in an iron-based superconductor

Majorana zero-modes (MZMs) are spatially-localized zero-energy fractional quasiparticles with non-Abelian braiding statistics that hold a great promise for topological quantum computing. Due to its particle-antiparticle equivalence, an MZM exhibits robust resonant Andreev reflection and 2e2/h quantized conductance at low temperature. By utilizing variable-tunnel-coupled scanning tunneling spectroscopy, we study tunneling conductance of vortex bound states on FeTe0.55Se0.45 superconductors. We report observations of conductance plateaus as a function of tunnel coupling for zero-energy vortex bound states with values close to or even reaching the 2e2/h quantum conductance. In contrast, no such plateau behaviors were observed on either finite energy Caroli-de Genne-Matricon bound states or in the continuum of electronic states outside the superconducting gap. This unique behavior of the zero-mode conductance reaching a plateau strongly supports the existence of MZMs in this iron-based superconductor, which serves as a promising single-material platform for Majorana braiding at a relatively high temperature.

cond-mat.supr-con↗

An Iterative Polishing Framework based on Quality Aware Masked Language Model for Chinese Poetry Generation

Owing to its unique literal and aesthetical characteristics, automatic generation of Chinese poetry is still challenging in Artificial Intelligence, which can hardly be straightforwardly realized by end-to-end methods. In this paper, we propose a novel iterative polishing framework for highly qualified Chinese poetry generation. In the first stage, an encoder-decoder structure is utilized to generate a poem draft. Afterwards, our proposed Quality-Aware Masked Language Model (QAMLM) is employed to polish the draft towards higher quality in terms of linguistics and literalness. Based on a multi-task learning scheme, QA-MLM is able to determine whether polishing is needed based on the poem draft. Furthermore, QAMLM is able to localize improper characters of the poem draft and substitute with newly predicted ones accordingly. Benefited from the masked language model structure, QAMLM incorporates global context information into the polishing process, which can obtain more appropriate polishing results than the unidirectional sequential decoding. Moreover, the iterative polishing process will be terminated automatically when QA-MLM regards the processed poem as a qualified one. Both human and automatic evaluation have been conducted, and the results demonstrate that our approach is effective to improve the performance of encoder-decoder structure.

cs.CL↗

UER: An Open-Source Toolkit for Pre-training Models

Existing works, including ELMO and BERT, have revealed the importance of pre-training for NLP tasks. While there does not exist a single pre-training model that works best in all cases, it is of necessity to develop a framework that is able to deploy various pre-training models efficiently. For this purpose, we propose an assemble-on-demand pre-training toolkit, namely Universal Encoder Representations (UER). UER is loosely coupled, and encapsulated with rich modules. By assembling modules on demand, users can either reproduce a state-of-the-art pre-training model or develop a pre-training model that remains unexplored. With UER, we have built a model zoo, which contains pre-trained models based on different corpora, encoders, and targets (objectives). With proper pre-trained models, we could achieve new state-of-the-art results on a range of downstream datasets.

cs.CL↗

PDANet: Polarity-consistent Deep Attention Network for Fine-grained Visual Emotion Regression

Existing methods on visual emotion analysis mainly focus on coarse-grained emotion classification, i.e. assigning an image with a dominant discrete emotion category. However, these methods cannot well reflect the complexity and subtlety of emotions. In this paper, we study the fine-grained regression problem of visual emotions based on convolutional neural networks (CNNs). Specifically, we develop a Polarity-consistent Deep Attention Network (PDANet), a novel network architecture that integrates attention into a CNN with an emotion polarity constraint. First, we propose to incorporate both spatial and channel-wise attentions into a CNN for visual emotion regression, which jointly considers the local spatial connectivity patterns along each channel and the interdependency between different channels. Second, we design a novel regression loss, i.e. polarity-consistent regression (PCR) loss, based on the weakly supervised emotion polarity to guide the attention generation. By optimizing the PCR loss, PDANet can generate a polarity preserved attention map and thus improve the emotion regression performance. Extensive experiments are conducted on the IAPS, NAPS, and EMOTIC datasets, and the results demonstrate that the proposed PDANet outperforms the state-of-the-art approaches by a large margin for fine-grained visual emotion regression. Our source code is released at: https://github.com/ZizhouJia/PDANet.

cs.CV↗

Bifrequency 3D Ghost Imaging with Haar Wavelet Transform

Recently, ghost imaging has been attracting attentions because its mechanism would lead to many applications inaccessible to conventional imaging methods. However, it is challenging for high contrast and high resolution imaging, due to its low signal-to-noise ratio (SNR) and the demand of high sampling rate in detection. To circumvent these challenges, we here propose a ghost imaging scheme that exploits Haar wavelets as illuminating patterns with a bi-frequency light projecting system and frequency-selecting single-pixel detectors. This method provides a theoretically 100% image contrast and high detection SNR, which reduces the requirement of high dynamic range of detectors, enabling high resolution ghost imaging. Moreover, it can highly reduce the sampling rate (far below Nyquist limit) for a sparse object by adaptively abandoning unnecessary patterns during the measurement. These characteristics are experimentally verified with a resolution of 512 times 512 and a sampling rate lower than 5%. A high-resolution (1000 times 1000 times 1000) 3D reconstruction of an object is also achieved from multi-angle images.

eess.IV↗

Modeling Hierarchical Usage Context for Software Exceptions based on Interaction Data

Traces of user interactions with a software system, captured in production, are commonly used as an input source for user experience testing. In this paper, we present an alternative use, introducing a novel approach of modeling user interaction traces enriched with another type of data gathered in production - software fault reports consisting of software exceptions and stack traces. The model described in this paper aims to improve developers' comprehension of the circumstances surrounding a specific software exception and can highlight specific user behaviors that lead to a high frequency of software faults. Modeling the combination of interaction traces and software crash reports to form an interpretable and useful model is challenging due to the complexity and variance in the combined data source. Therefore, we propose a probabilistic unsupervised learning approach, adapting the Nested Hierarchical Dirichlet Process, which is a Bayesian non-parametric topic model commonly applied to natural language data. This model infers a tree of topics, each of whom describes a set of commonly co-occurring commands and exceptions. The topic tree can be interpreted hierarchically to aid in categorizing the numerous types of exceptions and interactions. We apply the proposed approach to large scale datasets collected from the ABB RobotStudio software application, and evaluate it both numerically and with a small survey of the RobotStudio developers.

cs.SE↗

GRN: Gated Relation Network to Enhance Convolutional Neural Network for Named Entity Recognition

The dominant approaches for named entity recognition (NER) mostly adopt complex recurrent neural networks (RNN), e.g., long-short-term-memory (LSTM). However, RNNs are limited by their recurrent nature in terms of computational efficiency. In contrast, convolutional neural networks (CNN) can fully exploit the GPU parallelism with their feedforward architectures. However, little attention has been paid to performing NER with CNNs, mainly owing to their difficulties in capturing the long-term context information in a sequence. In this paper, we propose a simple but effective CNN-based network for NER, i.e., gated relation network (GRN), which is more capable than common CNNs in capturing long-term context. Specifically, in GRN we firstly employ CNNs to explore the local context features of each word. Then we model the relations between words and use them as gates to fuse local context features into global ones for predicting labels. Without using recurrent layers that process a sentence in a sequential manner, our GRN allows computations to be performed in parallel across the entire sentence. Experiments on two benchmark NER datasets (i.e., CoNLL2003 and Ontonotes 5.0) show that, our proposed GRN can achieve state-of-the-art performance with or without external knowledge. It also enjoys lower time costs to train and test.We have made the code publicly available at https://github.com/HuiChen24/NER-GRN.

cs.CL↗

A new Majorana platform in an Fe-As bilayer superconductor

Recently, iron-chalcogenide superconductors have emerged as a new and promising platform for studying and manipulating Majorana zero mode (MZM). By combining topological band structure and superconductivity in a multiband material, they provide significant advantages such as higher superconducting transition temperature (Tc) and isolated Majorana mode. However, iron-chalcogenide superconductors, especially Fe(Te,Se), suffer from strong inhomogeneity which may hamper their practical application. On the other hand, some iron-pnictide (Fe-As) superconductors, such as LiFeAs, have been demonstrated to have a similar topological band structure, yet no MZM has been observed in its vortex cores, raising a question of universality of MZM presence in iron-based superconductors. In this work, by using high-resolution angle-resolved photoemission spectroscopy and scanning tunneling microscopy/spectroscopy, we identify the first Fe-As superconductor CaKFe4As4 (Tc = 35 K) which has both the superconducting Dirac surface states and the MZMs inside its vortex cores. The topological band inversion is largely due to the down-shift of the pz band caused by the bilayer band folding in this material. More strikingly, the energies and spatial line profiles of MZM and multiple quantized Caroli-de Gennes-Matricon bound states observed inside the topological vortex can be accurately reproduced by a simple theoretical model derived from a surface Dirac cone, firmly establishing Majorana nature of the zero mode.

cond-mat.supr-con↗

Ultra-high-speed color imaging with single-pixel detectors under low light level

Single-pixel imaging is suitable for low light level scenarios because a bucket detector is employed to maximally collect the light from an object. However, one of the challenges is its slow imaging speed, mainly due to the slow light modulation technique. We here demonstrate 1.4MHz video imaging based on computational ghost imaging with a RGB LED array having a full-range frame rate up to 100MHz. With this method, the motion of a high speed propeller is observed. Moreover, by exploiting single-photon detectors to increase the detection efficiency, this method is developed for ultra-high-speed imaging under low light level.

physics.ins-det↗

Naked-Eye Ghost Imaging via Photoelectric-Feedback

Based on optical correlations, ghost imaging is usually reconstructed by computer algorithm from the acquired data. We here proposed an alternatively high contrast naked-eye ghost imaging scheme which avoids computer algorithm processing. Instead, the proposed scheme uses a photoelectric feedback loop to realize the multiplication process of traditional ghost imaging. Meanwhile, it exploits the vision persistence effect to implement integral process and to generate negative images observed by naked eyes. To realize high contrast naked-eye ghost imaging, a special pattern-scanning architecture on a low-speed light-modulation mask is designed, which enables high-resolution imaging with lower-order Hadamard vectors and boosts the imaging speed as well. Moreover, two kinds of feedback circuits, the digital circuit and the analog circuit, are presented respectively, which can achieve high-speed feedback operation on the light intensity. With this approach, we demonstrate high-contrast real-time imaging for moving objects.

eess.IV↗

All-optical naked-eye ghost imaging

Ghost imaging is usually based on optoelectronic process and eletronic computing. We here propose a new ghost imaging scheme, which avoids any optoelectronic or electronic process. Instead, the proposed scheme exploits all-optical correlation via the light-light interaction and the vision persistence effect to generate images observed by naked eyes. To realize high contrast naked-eye ghost imaging, a special pattern-scanning architecture on a low-speed light-modulation disk is designed, which also enables high-resolution imaging with lower-order Hadamard vectors and boosts the imaging speed. With this approach, we realize high-contrast real-time ghost imaging for moving colored objects.

eess.IV↗

Observation of half-integer level shift of vortex bound states in an iron-based superconductor

Vortices in topological superconductors host Majorana zero modes (MZMs), which are proposed to be building blocks of fault-tolerant topological quantum computers. Recently, a new single-material platform for realizing MZM has been discovered in iron-based superconductors, without involving hybrid semiconductor-superconductor structures. Here we report on a detailed scanning tunneling spectroscopy study of a FeTe0.55Se0.45 single crystal, revealing two distinct classes of vortices present in this system which differ by a half-integer level shift in the energy spectra of the vortex bound states. This level shift is directly tied with the presence or absence of zero-bias peak and also alters the ratios of higher energy levels from integer to half-odd-integer. Our model calculations fully reproduce the spectra of these two types of vortex bound states, suggesting the presence of topological and conventional superconducting regions that coexist within the same crystal. Our findings provide strong evidence for the topological nature of superconductivity in FeTe0.55Se0.45 and establish it as an excellent platform for further studies on MZMs.

cond-mat.supr-con↗