arXiv subjects

Jian Liu

Publications and source records attributed to Jian Liu.

At least 361 records · Page 20Linked to original sources

arXiv2021

On-chip multi-color microdisk laser on Yb3+-doped thin-film lithium niobate

We demonstrate an on-chip Yb3+-doped lithium niobate (LN) microdisk laser. The intrinsic quality factors of the fabricated Yb3+-doped LN microdisk resonator are measured up to 3.79x10^5 at 976 nm wavelength and 1.1x10^6 at 1514 nm wavelength. The multi-mode laser emissions are obtained in a band from 1020 nm to 1070 nm pumped by 984 nm laser and with the low threshold of 103 μW, resulting in a slope efficiency of 0.53% at room temperature. Furthermore, the second-harmonic frequency of pump light and the sum-frequency of the pump light and laser emissions are both generated in the on-chip Yb3+-doped LN microdisk benefited from the strong \c{hi}(2) nonlinearity of LN. These microdisk lasers are expected to contribute to the high-density integration of LNOI-based photonic chip.

physics.optics↗

arXiv2021

Learn to Forget: Machine Unlearning via Neuron Masking

Nowadays, machine learning models, especially neural networks, become prevalent in many real-world applications.These models are trained based on a one-way trip from user data: as long as users contribute their data, there is no way to withdraw; and it is well-known that a neural network memorizes its training data. This contradicts the "right to be forgotten" clause of GDPR, potentially leading to law violations. To this end, machine unlearning becomes a popular research topic, which allows users to eliminate memorization of their private data from a trained machine learning model.In this paper, we propose the first uniform metric called for-getting rate to measure the effectiveness of a machine unlearning method. It is based on the concept of membership inference and describes the transformation rate of the eliminated data from "memorized" to "unknown" after conducting unlearning. We also propose a novel unlearning method calledForsaken. It is superior to previous work in either utility or efficiency (when achieving the same forgetting rate). We benchmark Forsaken with eight standard datasets to evaluate its performance. The experimental results show that it can achieve more than 90\% forgetting rate on average and only causeless than 5\% accuracy loss.

cs.LG↗

arXiv2021

Commutator Matrix in Phase Space Mapping Models for Nonadiabatic Quantum Dynamics

We show that a novel, general phase space mapping Hamiltonian for nonadiabatic systems, which is reminiscent of the renowned Meyer-Miller mapping Hamiltonian, involves a commutator variable matrix rather than the conventional zero-point-energy parameter. In the exact mapping formulation on constraint space for phase space approaches for nonadiabatic dynamics, the general mapping Hamiltonian with commutator variables can be employed to generate approximate trajectory-based dynamics. Various benchmark model tests, which range from gas phase to condensed phase systems, suggest that the overall performance of the general mapping Hamiltonian is better than that of the conventional Meyer-Miller Hamiltonian.

physics.chem-ph↗

arXiv2021

Proposal for constraining non-Newtonian gravity at nm range via criticality enhanced measurement of resonance frequency shift

We propose a quantum mechanical method of constraining non-Newtonian gravity at the nanometer range. In this method, a hybrid electro-optomechanical system is employed. Applying a strong driving field, we can obtain normal mode splitting of the electromechanical subsystem which is related to the resonance frequency of the mechanical oscillator. Moreover, we investigate the relationship between the variance of normal mode splitting and the resonance frequency shift induced by the gradient of exotic forces provided that our system is operated at critical points. Furthermore, via suppressing the Casimir background, we set a constraint on the non-Newtonian gravity which improves the previous bounds by about a factor of 7 at 1 nanometer range. Our results indicate that our method could be put into consideration in relevant experimental searches.

quant-ph↗

arXiv2021

Proposal for measuring Newtonian constant of gravitation at an exceptional point in an optomechanical system

We develop a quantum mechanical method of measuring the Newtonian constant of gravitation, G. In this method, an optomechanical system consisting of two cavities and two membrane resonators is used. The added source mass would induce the shifts of the eigenfrequencies of the supermodes. Via detecting the shifts, we can perform our measurement of G. Furthermore, our system can features exceptional point (EP) which are branch point singularities of the spectrum and eigenfunctions. In the paper, we demonstrate that operating the system at EP can enhance our measurement of G. In addition, we derive the relationship between EP enlarged eigenfrequency shift and the Newtonian constant. This work provides a way to engineer EP-assisted optomechanical devices for applications in the field of precision measurement of G

quant-ph↗

arXiv2021

Constraining Axion-to-Nucleon interaction via ultranarrow linewidth in the Casimir-less regime

In this paper we develop a quantum optical method to detect the axion-nucleon interaction. We ultilize a levitated optomechanical system consisting of a silica nanosphere and an optical cavity here. We translate the trapping positions of the nanosphere, resulting the shift of its resonance frequency, which can be determined from measuring the resulting resonance shift in the transmission spectrum. Furthermore, The frequency shift can be related to the additional forces due to two-axion exchange via substraction. Based on noise ananlysis, estimation and calculation, we set the stringent prospective constraints for the coupling constants of axion-neucleon interaction $g_{an}$ and $g_{ap}$. In the case of $g_{an}^2 = g_{ap}^2$ , our constraints are most stringent at an ultrawide axion mass range approximately from $10^{-4}μeV$ to $10$ $eV$ .

quant-ph↗

arXiv2021

Template-Based Named Entity Recognition Using BART

There is a recent interest in investigating few-shot NER, where the low-resource target domain has different label sets compared with a resource-rich source domain. Existing methods use a similarity-based metric. However, they cannot make full use of knowledge transfer in NER model parameters. To address the issue, we propose a template-based method for NER, treating NER as a language model ranking problem in a sequence-to-sequence framework, where original sentences and statement templates filled by candidate named entity span are regarded as the source sequence and the target sequence, respectively. For inference, the model is required to classify each candidate span based on the corresponding template scores. Our experiments demonstrate that the proposed method achieves 92.55% F1 score on the CoNLL03 (rich-resource task), and significantly better than fine-tuning BERT 10.88%, 15.34%, and 11.73% F1 score on the MIT Movie, the MIT Restaurant, and the ATIS (low-resource task), respectively.

cs.CL↗

arXiv2021

The transport-structural correspondence across the nematic phase transition probed by elasto-x-ray diffraction

Electronic nematicity in iron pnictide materials is coupled to both the lattice and the conducting electrons, which allows both structural and transport observables to probe nematic fluctuations and the order parameter. Here we combine simultaneous transport and x-ray diffraction measurements with in-situ tunable strain (elasto-XRD) to measure the temperature dependence of the shear modulus and elastoresistivity above the nematic transition and the spontaneous orthorhombicity and resistivity anisotropy below the nematic transition, all within a single sample of $Ba(Fe_{0.96}Co_{0.04})_{2} As_{2}$. The ratio of transport to structural quantities is nearly temperature-independent over a 74 K range and agrees between the ordered and disordered phases. These results show that elasto-XRD is a powerful technique to probe the nemato-elastic and nemato-transport couplings, which have important implications to the nearby superconductivity. It also enables the measurement in the large strain limit, where the breakdown of mean field description reveals the intertwined nature of nematicity.

cond-mat.str-el↗

arXiv2021

Strongly anisotropic antiferromagnetic coupling in EuFe2As2 revealed by stress detwinning

Of all parent compounds of iron-based high-temperature superconductors, EuFe2As2 exhibits by far the largest magnetostructural coupling due to the sizable biquadratic interaction between Eu and Fe moments. While the coupling between Eu antiferromagnetic order and Fe structural/antiferromagnetic domains enables rapid field detwinning, this prevents simple magnetometry measurements from extracting the critical fields of the Eu metamagnetic transition. Here we measure these critical fields by combining x-ray magnetic circular dichroism spectroscopy with in-situ tunable uniaxial stress and applied magnetic field. The combination of two tuning knobs allows us to separate the stress-detwinning of structural domains from the field-induced reorientation of Eu moments. Intriguingly, we find a spin-flip transition which can only result from a strongly anisotropic interaction between Eu planes. We argue that this anisotropic exchange is a consequence of the strong anisotropy in the magnetically ordered Fe layer, which presents a new form of higher-order coupling between Eu and Fe magnetism.

cond-mat.supr-con↗

arXiv2021

OFEI: A Semi-black-box Android Adversarial Sample Attack Framework Against DLaaS

With the growing popularity of Android devices, Android malware is seriously threatening the safety of users. Although such threats can be detected by deep learning as a service (DLaaS), deep neural networks as the weakest part of DLaaS are often deceived by the adversarial samples elaborated by attackers. In this paper, we propose a new semi-black-box attack framework called one-feature-each-iteration (OFEI) to craft Android adversarial samples. This framework modifies as few features as possible and requires less classifier information to fool the classifier. We conduct a controlled experiment to evaluate our OFEI framework by comparing it with the benchmark methods JSMF, GenAttack and pointwise attack. The experimental results show that our OFEI has a higher misclassification rate of 98.25%. Furthermore, OFEI can extend the traditional white-box attack methods in the image field, such as fast gradient sign method (FGSM) and DeepFool, to craft adversarial samples for Android. Finally, to enhance the security of DLaaS, we use two uncertainties of the Bayesian neural network to construct the combined uncertainty, which is used to detect adversarial samples and achieves a high detection rate of 99.28%.

cs.CR↗

arXiv2021

Unsupervised Sentiment Analysis by Transferring Multi-source Knowledge

Sentiment analysis (SA) is an important research area in cognitive computation-thus in-depth studies of patterns of sentiment analysis are necessary. At present, rich resource data-based SA has been well developed, while the more challenging and practical multi-source unsupervised SA (i.e. a target domain SA by transferring from multiple source domains) is seldom studied. The challenges behind this problem mainly locate in the lack of supervision information, the semantic gaps among domains (i.e., domain shifts), and the loss of knowledge. However, existing methods either lack the distinguishable capacity of the semantic gaps among domains or lose private knowledge. To alleviate these problems, we propose a two-stage domain adaptation framework. In the first stage, a multi-task methodology-based shared-private architecture is employed to explicitly model the domain common features and the domain-specific features for the labeled source domains. In the second stage, two elaborate mechanisms are embedded in the shared private architecture to transfer knowledge from multiple source domains. The first mechanism is a selective domain adaptation (SDA) method, which transfers knowledge from the closest source domain. And the second mechanism is a target-oriented ensemble (TOE) method, in which knowledge is transferred through a well-designed ensemble method. Extensive experiment evaluations verify that the performance of the proposed framework outperforms unsupervised state-of-the-art competitors. What can be concluded from the experiments is that transferring from very different distributed source domains may degrade the target-domain performance, and it is crucial to choose the proper source domains to transfer from.

cs.CL↗

arXiv2021

High Order Explicit Lorentz Invariant Volume-preserving Algorithms for Relativistic Dynamics of Charged Particles

Lorentz invariant structure-preserving algorithms possess reference-independent secular stability, which is vital for simulating relativistic multi-scale dynamical processes. The splitting method has been widely used to construct structure-preserving algorithms, but without exquisite considerations, it can easily break the Lorentz invariance of continuous systems. In this paper, we introduce a Lorentz invariant splitting technique to construct high order explicit Lorentz invariant volume-preserving algorithms (LIVPAs) for simulating charged particle dynamics. Using this method, long-term stable explicit LIVPAs with different orders are developed and their performances of Lorentz invariance and long-term stability are analyzed in detail.

physics.plasm-ph↗

arXiv2021

Manipulate the Electronic State of Mott Iridate Superlattice through Protonation Induced Electron-Filling

Spin-orbit-coupled Mott iridates show great similarity with parent compounds of superconducting cuprates, attracting extensive research interests especially for their electron-doped states. However, previous experiments are largely limited within a small doping range due to the absence of effective dopants, and therefore the electron-doped phase diagram remains elusive. Here we utilize an ionic-liquid-gating induced protonation method to achieve electron-doping into a 5d Mott-insulator built with SrIrO3/SrTiO3 superlattice, and achieve a systematic mapping of its electron-doped phase diagram with the evolution of the iridium valence state from 4+ to 3+, equivalent to doping of one electron per iridium ion. Along increasing doping level, the parent Mott-insulator is first turned into a localized metallic state with gradually suppressed magnetic ordering, and then further evolved into a nonmagnetic band insulating state. This work forms an important step forward for the study of electron-doped Mott iridate systems, and the strategy of manipulating the band filling in an artificially designed superlattice structure can be readily extended into other systems with more exotic states to explore.

cond-mat.str-el↗

arXiv2021

CelebA-Spoof Challenge 2020 on Face Anti-Spoofing: Methods and Results

As facial interaction systems are prevalently deployed, security and reliability of these systems become a critical issue, with substantial research efforts devoted. Among them, face anti-spoofing emerges as an important area, whose objective is to identify whether a presented face is live or spoof. Recently, a large-scale face anti-spoofing dataset, CelebA-Spoof which comprised of 625,537 pictures of 10,177 subjects has been released. It is the largest face anti-spoofing dataset in terms of the numbers of the data and the subjects. This paper reports methods and results in the CelebA-Spoof Challenge 2020 on Face AntiSpoofing which employs the CelebA-Spoof dataset. The model evaluation is conducted online on the hidden test set. A total of 134 participants registered for the competition, and 19 teams made valid submissions. We will analyze the top ranked solutions and present some discussion on future work directions.

cs.CV↗

arXiv2021

A Künneth Formula of Hypergraphs

In this paper, based on the embedded homology groups of hypergraphs defined in \cite{h1}, we define the product of hypergraphs and prove the corresponding Künneth formula of hypergraphs which can be generalized to the Künneth formula for the embedded homology of graded subsets of chain complexes with coefficients in a principal ideal domain.

math.AT↗

arXiv2021

Enabling Fast and Universal Audio Adversarial Attack Using Generative Model

Recently, the vulnerability of DNN-based audio systems to adversarial attacks has obtained the increasing attention. However, the existing audio adversarial attacks allow the adversary to possess the entire user's audio input as well as granting sufficient time budget to generate the adversarial perturbations. These idealized assumptions, however, makes the existing audio adversarial attacks mostly impossible to be launched in a timely fashion in practice (e.g., playing unnoticeable adversarial perturbations along with user's streaming input). To overcome these limitations, in this paper we propose fast audio adversarial perturbation generator (FAPG), which uses generative model to generate adversarial perturbations for the audio input in a single forward pass, thereby drastically improving the perturbation generation speed. Built on the top of FAPG, we further propose universal audio adversarial perturbation generator (UAPG), a scheme crafting universal adversarial perturbation that can be imposed on arbitrary benign audio input to cause misclassification. Extensive experiments show that our proposed FAPG can achieve up to 167X speedup over the state-of-the-art audio adversarial attack methods. Also our proposed UAPG can generate universal adversarial perturbation that achieves much better attack performance than the state-of-the-art solutions.

eess.AS↗

arXiv2021

A partial converse ghost lemma for the derived category of a commutative noetherian ring

In this article a condition is given to detect the containment among thick subcategories of the bounded derived category of a commutative noetherian ring. More precisely, for a commutative noetherian ring $R$ and complexes of $R$-modules with finitely generated homology $M$ and $N$, we show $N$ is in the thick subcategory generated by $M$ if and only if the ghost index of $N_\mathfrak{p}$ with respect to $M_\mathfrak{p}$ is finite for each prime $\mathfrak{p}$ of $R$. To do so, we establish a "converse coghost lemma" for the bounded derived category of a non-negatively graded DG algebra with noetherian homology.

math.AC↗

arXiv2020

A hybrid deep-learning approach for complex biochemical named entity recognition

Named entity recognition (NER) of chemicals and drugs is a critical domain of information extraction in biochemical research. NER provides support for text mining in biochemical reactions, including entity relation extraction, attribute extraction, and metabolic response relationship extraction. However, the existence of complex naming characteristics in the biomedical field, such as polysemy and special characters, make the NER task very challenging. Here, we propose a hybrid deep learning approach to improve the recognition accuracy of NER. Specifically, our approach applies the Bidirectional Encoder Representations from Transformers (BERT) model to extract the underlying features of the text, learns a representation of the context of the text through Bi-directional Long Short-Term Memory (BILSTM), and incorporates the multi-head attention (MHATT) mechanism to extract chapter-level features. In this approach, the MHATT mechanism aims to improve the recognition accuracy of abbreviations to efficiently deal with the problem of inconsistency in full-text labels. Moreover, conditional random field (CRF) is used to label sequence tags because this probabilistic method does not need strict independence assumptions and can accommodate arbitrary context information. The experimental evaluation on a publicly-available dataset shows that the proposed hybrid approach achieves the best recognition performance; in particular, it substantially improves performance in recognizing abbreviations, polysemes, and low-frequency entities, compared with the state-of-the-art approaches. For instance, compared with the recognition accuracies for low-frequency entities produced by the BILSTM-CRF algorithm, those produced by the hybrid approach on two entity datasets (MULTIPLE and IDENTIFIER) have been increased by 80% and 21.69%, respectively.

cs.CL↗