Searcharxiv⌕ Search

arXiv subjects

Mei Wang

Publications and source records attributed to Mei Wang.

At least 55 records · Page 3Linked to original sources
arXiv2021

Predicting Disease Progress with Imprecise Lab Test Results

In existing deep learning methods, almost all loss functions assume that sample data values used to be predicted are the only correct ones. This assumption does not hold for laboratory test data. Test results are often within tolerable or imprecision ranges, with all values in the ranges acceptable. By considering imprecision samples, we propose an imprecision range loss (IR loss) method and incorporate it into Long Short Term Memory (LSTM) model for disease progress prediction. In this method, each sample in imprecision range space has a certain probability to be the real value, participating in the loss calculation. The loss is defined as the integral of the error of each point in the impression range space. A sampling method for imprecision space is formulated. The continuous imprecision space is discretized, and a sequence of imprecise data sets are obtained, which is convenient for gradient descent learning. A heuristic learning algorithm is developed to learn the model parameters based on the imprecise data sets. Experimental results on real data show that the prediction method based on IR loss can provide more stable and consistent prediction result when test samples are generated from imprecision range.

cs.LG↗
arXiv2021

Nonreciprocal high-order sidebands induced by magnon Kerr nonlinearity

We propose an effective approach for creating robust nonreciprocity of high-order sidebands, including the first-, second- and third-order sidebands, at microwave frequencies. This approach relies on magnon Kerr nonlinearity in a cavity magnonics system composed of two microwave cavities and one yttrium iron garnet (YIG) sphere. By manipulating the driving power applied on YIG and the frequency detuning between the magnon mode in YIG and the driving field, the effective Kerr nonlinearity can be strengthened, thereby inducing strong transmission non-reciprocity. More interestingly, we find the higher the sideband order, the stronger the transmission nonreciprocity marked by the higher isolation ratio in the optimal detuning regime. Such a series of equally-spaced high-order sidebands have potential applications in frequency comb-like precision measurement, besides structuring high-performance on-chip nonreciprocal devices.

physics.app-ph↗
arXiv2020

A hidden anti-jamming method based on deep reinforcement learning

Most of the current anti-jamming algorithms for wireless communications only consider how to avoid jamming attacks, but ignore that the communication waveform or frequency action may be obtained by the jammers. Although existing anti-jamming methods can guarantee temporary communication effects, the long-term performance of these anti-jamming methods may be depressed when intelligent jammers are capable of learning from historical communication activities. Aiming at this issue, a hidden anti-jamming method based on the idea of reducing the jammer's sense probability is proposed. Firstly, the sensing probability of the jammer is obtained by calculating the correlation between the actions of the jammer and the user. Later, a deep reinforcement learning framework is designed, which aims at not only maximizing the communication throughput but also minimizing the action's correlation between the jammer and the user. Finally, a hidden anti-jamming algorithm is proposed, which links the instantaneous return with the communication quality of users and the correlation between users and jammer. The simulation result shows that the proposed algorithm not only avoids being sensed by the jammer but also improves its anti-jamming performance compared to the current algorithm that only considers jamming avoidance.

cs.IT↗
arXiv2020

Deep Face Recognition: A Survey

Deep learning applies multiple processing layers to learn representations of data with multiple levels of feature extraction. This emerging technique has reshaped the research landscape of face recognition (FR) since 2014, launched by the breakthroughs of DeepFace and DeepID. Since then, deep learning technique, characterized by the hierarchical architecture to stitch together pixels into invariant face representation, has dramatically improved the state-of-the-art performance and fostered successful real-world applications. In this survey, we provide a comprehensive review of the recent developments on deep FR, covering broad topics on algorithm designs, databases, protocols, and application scenes. First, we summarize different network architectures and loss functions proposed in the rapid evolution of the deep FR methods. Second, the related face processing methods are categorized into two classes: "one-to-many augmentation" and "many-to-one normalization". Then, we summarize and compare the commonly used databases for both model training and evaluation. Third, we review miscellaneous scenes in deep FR, such as cross-factor, heterogenous, multiple-media and industrial scenes. Finally, the technical challenges and several promising directions are highlighted.

cs.CV↗
arXiv2020

Impact of Medical Data Imprecision on Learning Results

Test data measured by medical instruments often carry imprecise ranges that include the true values. The latter are not obtainable in virtually all cases. Most learning algorithms, however, carry out arithmetical calculations that are subject to uncertain influence in both the learning process to obtain models and applications of the learned models in, e.g. prediction. In this paper, we initiate a study on the impact of imprecision on prediction results in a healthcare application where a pre-trained model is used to predict future state of hyperthyroidism for patients. We formulate a model for data imprecisions. Using parameters to control the degree of imprecision, imprecise samples for comparison experiments can be generated using this model. Further, a group of measures are defined to evaluate the different impacts quantitatively. More specifically, the statistics to measure the inconsistent prediction for individual patients are defined. We perform experimental evaluations to compare prediction results based on the data from the original dataset and the corresponding ones generated from the proposed precision model using the long-short-term memories (LSTM) network. The results against a real world hyperthyroidism dataset provide insights into how small imprecisions can cause large ranges of predicted results, which could cause mis-labeling and inappropriate actions (treatments or no treatments) for individual patients.

cs.LG↗
arXiv2019

Mitigate Bias in Face Recognition using Skewness-Aware Reinforcement Learning

Racial equality is an important theme of international human rights law, but it has been largely obscured when the overall face recognition accuracy is pursued blindly. More facts indicate racial bias indeed degrades the fairness of recognition system and the error rates on non-Caucasians are usually much higher than Caucasians. To encourage fairness, we introduce the idea of adaptive margin to learn balanced performance for different races based on large margin losses. A reinforcement learning based race balance network (RL-RBN) is proposed. We formulate the process of finding the optimal margins for non-Caucasians as a Markov decision process and employ deep Q-learning to learn policies for an agent to select appropriate margin by approximating the Q-value function. Guided by the agent, the skewness of feature scatter between races can be reduced. Besides, we provide two ethnicity aware training datasets, called BUPT-Globalface and BUPT-Balancedface dataset, which can be utilized to study racial bias from both data and algorithm aspects. Extensive experiments on RFW database show that RL-RBN successfully mitigates racial bias and learns more balanced performance for different races.

cs.CV↗
arXiv2019

Time-weighted Attentional Session-Aware Recommender System

Session-based Recurrent Neural Networks (RNNs) are gaining increasing popularity for recommendation task, due to the high autocorrelation of user's behavior on the latest session and the effectiveness of RNN to capture the sequence order information. However, most existing session-based RNN recommender systems still solely focus on the short-term interactions within a single session and completely discard all the other long-term data across different sessions. While traditional Collaborative Filtering (CF) methods have many advanced research works on exploring long-term dependency, which show great value to be explored and exploited in deep learning models. Therefore, in this paper, we propose ASARS, a novel framework that effectively imports the temporal dynamics methodology in CF into session-based RNN system in DL, such that the temporal info can act as scalable weights by a parallel attentional network. Specifically, we first conduct an extensive data analysis to show the distribution and importance of such temporal interactions data both within sessions and across sessions. And then, our ASARS framework promotes two novel models: (1) an inter-session temporal dynamic model that captures the long-term user interaction for RNN recommender system. We integrate the time changes in session RNN and add user preferences as model drifting; and (2) a novel triangle parallel attention network that enhances the original RNN model by incorporating time information. Such triangle parallel network is also specially designed for realizing data argumentation in sequence-to-scalar RNN architecture, and thus it can be trained very efficiently. Our extensive experiments on four real datasets from different domains demonstrate the effectiveness and large improvement of ASARS for personalized recommendation.

cs.LG↗
arXiv2019

Racial Faces in-the-Wild: Reducing Racial Bias by Information Maximization Adaptation Network

Racial bias is an important issue in biometric, but has not been thoroughly studied in deep face recognition. In this paper, we first contribute a dedicated dataset called Racial Faces in-the-Wild (RFW) database, on which we firmly validated the racial bias of four commercial APIs and four state-of-the-art (SOTA) algorithms. Then, we further present the solution using deep unsupervised domain adaptation and propose a deep information maximization adaptation network (IMAN) to alleviate this bias by using Caucasian as source domain and other races as target domains. This unsupervised method simultaneously aligns global distribution to decrease race gap at domain-level, and learns the discriminative target representations at cluster level. A novel mutual information loss is proposed to further enhance the discriminative ability of network output without label information. Extensive experiments on RFW, GBU, and IJB-A databases show that IMAN successfully learns features that generalize well across different races and across different databases.

cs.CV↗
arXiv2019

Broadcast Age of Information in CSMA/CA Based Wireless Networks

We consider a wireless sensor network in which all the nodes wish to spread their updates over the network using CSMA/CA protocol. We investigate the age of information of the spreading process from a transmitter perspective, which is referred to as the \textit{broadcast age of information (BAoI)}. To be specific, BAoI is the age of the latest update successfully broadcasted to the one-hop neighbors of a node, and thus is suitable to measure the rapidity of the update spreading process. We establish an equivalent transmission model of the network by deriving the transmission probability and the collision probability of nodes. With this equivalent model, we then present the average BAoI of the network explicitly. Our results present the scaling laws of average BAoI with respect to node density and frame length, and are further illustrated through numerical results. As is shown, the average BAoI is increasing with node density and is convex in frame length, i.e., would be large when frame length is very small or very large.

cs.IT↗
arXiv2018

PT-Symmetric magnetic Chaos in cavity magnomechanics

Here, we research a novel cavity magnomechanical system, where magnon driven by a microwave field couples with a phonon mode with a nonlinear magneostrictive interaction (radiation pressure-like). Based on this interaction, we numerically demonstrate PT-symmetric chaos in this system. With only one monochrome driving, the chaotic threshold is lowered to a rather low levels, this is due to the dynamical enhancement of nonlinearity in the PT-symmetry broken phase. Moreover, by simply manipulating the phase transition between PT-symmetry phase and PT-symmetry broken phase, we can switch the system between into and out of chaotic regimes. Our work may broaden the cavity magnomechanics and provide a promising application for magnetic chaos-related security communications.

cond-mat.mes-hall↗
arXiv2018

Unconventional phonon blockade via atom-photon-phonon interaction in hybrid optomechanical systems

Phonon nonlinearities play an important role in hybrid quantum networks and on-chip quantum devices. We investigate the phonon statistics of a mechanical oscillator in hybrid systems composed of an atom and one or two standard optomechanical cavities. An efficiently enhanced atom-phonon interaction can be derived via a tripartite atom-photon-phonon interaction, where the atom-photon coupling depends on the mechanical displacement without practically changing a cavity frequency. This novel mechanism of optomechanical interactions, as predicted recently by Cotrufo et al. [Phys. Rev. Lett. 118, 133603 (2017)], is fundamentally different from standard ones. In the enhanced atom-phonon coupling, the strong phonon nonlinearity at a single-excitation level is obtained in the originally weak-coupling regime, which leads to the appearance of phonon blockade. Moreover, the optimal parameter regimes are presented both for the cases of one- and two- cavities. We compared phonon-number correlation functions of different orders for mechanical steady states generated in the one-cavity hybrid system, revealing the occurrence of phonon-induced tunneling and different types of phonon blockade. Our approach offers an alternative method to generate and control a single phonon in the quantum regime, and has potential applications in single-phonon quantum technologies.

quant-ph↗
arXiv2018

Constructing Social Networks From Binary Data

Much of applied network analysis concerns with studying the existing relationships between a set of agents; however, little focus has been given to the considerations of how to represent observed phenomena as a network object. In the case of physical structures such as electric grids or transportation flows, the construction of a network model is fairly straightforward as the nodes and edges usually correspond to some physical structure themselves. On the other hand, construction of a social network is much less defined; while nodes may correspond to well defined social agents, such as people or groups of people, there much more liberty in defining the relationship that edges should represent. This paper studies the intricacies of constructing a social network from data, in particular, binary data, applicable to a wide range of social science contexts. We examine several methods of constructing social networks in prior literature and discuss the methods under a common framework. Finally, using a data set of meetings among technologists in New York, we show that the different constructions of social networks arising from various interpretations of the underlying social relationships can result in vastly different network structures. These findings highlight the significance of understanding the precise relationships of interest when building a network model.

cs.SI↗
arXiv2018

Deep Visual Domain Adaptation: A Survey

Deep domain adaption has emerged as a new learning technique to address the lack of massive amounts of labeled data. Compared to conventional methods, which learn shared feature subspaces or reuse important source instances with shallow representations, deep domain adaption methods leverage deep networks to learn more transferable representations by embedding domain adaptation in the pipeline of deep learning. There have been comprehensive surveys for shallow domain adaption, but few timely reviews the emerging deep learning based methods. In this paper, we provide a comprehensive survey of deep domain adaptation methods for computer vision applications with four major contributions. First, we present a taxonomy of different deep domain adaption scenarios according to the properties of data that define how two domains are diverged. Second, we summarize deep domain adaption approaches into several categories based on training loss, and analyze and compare briefly the state-of-the-art methods under these categories. Third, we overview the computer vision applications that go beyond image classification, such as face recognition, semantic segmentation and object detection. Fourth, some potential deficiencies of current methods and several future directions are highlighted.

cs.CV↗
arXiv2017

Nonlinear effects in modulated quantum optomechanics

The nonlinear quantum regime is crucial for implementing interesting quantum effects, which have wide applications in modern quantum science. Here we propose an effective method to reach the nonlinear quantum regime in a modulated optomechanical system (OMS), which is originally in the weak-coupling regime. The mechanical spring constant and optomechanical interaction are modulated periodically. This leads to the result that the resonant optomechanical interaction can be effectively enhanced into the single-photon strong-coupling regime by the modulation-induced mechanical parametric amplification. Moreover, the amplified phonon noise can be suppressed completely by introducing a squeezed vacuum reservoir, which ultimately leads to the realization of photon blockade in a weakly coupled OMS. The reached nonlinear quantum regime also allows us to engineer the nonclassical states (e.g., Schrödinger cat states) of cavity field, which are robust against the phonon noise. This work offers an alternative approach to enhance the quantum nonlinearity of an OMS, which should expand the applications of cavity optomechanics in the quantum realm.

quant-ph↗
arXiv2017

Macroscopic quantum entanglement in modulated optomechanics

Quantum entanglement in mechanical systems is not only a key signature of macroscopic quantum effects, but has wide applications in quantum technologies. Here we proposed an effective approach for creating strong steady-state entanglement between two directly coupled mechanical oscillators (or a mechanical oscillator and a microwave resonator) in a modulated optomechanical system. The entanglement is achieved by combining the processes of a cavity cooling and the two-mode parametric interaction, which can surpass the bound on the maximal stationary entanglement from the two-mode parametric interaction. In principle, our proposal allows one to cool the system from an initial thermal state to an entangled state with high purity by a monochromatic driving laser. Also, the obtained entangled state can be used to implement the continuous-variable teleportation with high fidelity. Moreover, our proposal is robust against the thermal fluctuations of the mechanical modes under the condition of strong optical pumping.

quant-ph↗
arXiv2016

Testing gravity of a regular and slowly rotating phantom black hole by quasi-periodic oscillations

We extend firstly the regular phantom black hole solution to a slowly rotating black hole case and find that the phantom field depresses the angular velocity of the event horizon and suppresses the super-radiation of black hole. We also probe the dependence of quasi-periodic oscillations frequencies in relativistic precession model on the phantom parameter. With the observation data of GRO J1655-40, we make a constraint on the parameters of the regular and slowly rotating phantom black hole. Our results show that although the best-fit value of the phantom parameter $b$ is small, the allowed value of $b$ in the $1σ$ region is $b<0.619$, which means that the phantom theoretical model can not be excluded by the constraint from quasi-periodic oscillations with the observation data of GRO J1655-40.

gr-qc↗