arXiv subjects

Jingjing Wang

Publications and source records attributed to Jingjing Wang.

At least 109 records · Page 6Linked to original sources

arXiv2022

Semi-supervised Ranking for Object Image Blur Assessment

Assessing the blurriness of an object image is fundamentally important to improve the performance for object recognition and retrieval. The main challenge lies in the lack of abundant images with reliable labels and effective learning strategies. Current datasets are labeled with limited and confused quality levels. To overcome this limitation, we propose to label the rank relationships between pairwise images rather their quality levels, since it is much easier for humans to label, and establish a large-scale realistic face image blur assessment dataset with reliable labels. Based on this dataset, we propose a method to obtain the blur scores only with the pairwise rank labels as supervision. Moreover, to further improve the performance, we propose a self-supervised method based on quadruplet ranking consistency to leverage the unlabeled data more effectively. The supervised and self-supervised methods constitute a final semi-supervised learning framework, which can be trained end-to-end. Experimental results demonstrate the effectiveness of our method.

cs.CV↗

arXiv2022

Unimodal-Concentrated Loss: Fully Adaptive Label Distribution Learning for Ordinal Regression

Learning from a label distribution has achieved promising results on ordinal regression tasks such as facial age and head pose estimation wherein, the concept of adaptive label distribution learning (ALDL) has drawn lots of attention recently for its superiority in theory. However, compared with the methods assuming fixed form label distribution, ALDL methods have not achieved better performance. We argue that existing ALDL algorithms do not fully exploit the intrinsic properties of ordinal regression. In this paper, we emphatically summarize that learning an adaptive label distribution on ordinal regression tasks should follow three principles. First, the probability corresponding to the ground-truth should be the highest in label distribution. Second, the probabilities of neighboring labels should decrease with the increase of distance away from the ground-truth, i.e., the distribution is unimodal. Third, the label distribution should vary with samples changing, and even be distinct for different instances with the same label, due to the different levels of difficulty and ambiguity. Under the premise of these principles, we propose a novel loss function for fully adaptive label distribution learning, namely unimodal-concentrated loss. Specifically, the unimodal loss derived from the learning to rank strategy constrains the distribution to be unimodal. Furthermore, the estimation error and the variance of the predicted distribution for a specific sample are integrated into the proposed concentrated loss to make the predicted distribution maximize at the ground-truth and vary according to the predicting uncertainty. Extensive experimental results on typical ordinal regression tasks including age and head pose estimation, show the superiority of our proposed unimodal-concentrated loss compared with existing loss functions.

cs.CV↗

arXiv2022

Few-shot One-class Domain Adaptation Based on Frequency for Iris Presentation Attack Detection

Iris presentation attack detection (PAD) has achieved remarkable success to ensure the reliability and security of iris recognition systems. Most existing methods exploit discriminative features in the spatial domain and report outstanding performance under intra-dataset settings. However, the degradation of performance is inevitable under cross-dataset settings, suffering from domain shift. In consideration of real-world applications, a small number of bonafide samples are easily accessible. We thus define a new domain adaptation setting called Few-shot One-class Domain Adaptation (FODA), where adaptation only relies on a limited number of target bonafide samples. To address this problem, we propose a novel FODA framework based on the expressive power of frequency information. Specifically, our method integrates frequency-related information through two proposed modules. Frequency-based Attention Module (FAM) aggregates frequency information into spatial attention and explicitly emphasizes high-frequency fine-grained features. Frequency Mixing Module (FMM) mixes certain frequency components to generate large-scale target-style samples for adaptation with limited target bonafide samples. Extensive experiments on LivDet-Iris 2017 dataset demonstrate the proposed method achieves state-of-the-art or competitive performance under both cross-dataset and intra-dataset settings.

cs.CV↗

arXiv2022

Weakly Supervised Regional and Temporal Learning for Facial Action Unit Recognition

Automatic facial action unit (AU) recognition is a challenging task due to the scarcity of manual annotations. To alleviate this problem, a large amount of efforts has been dedicated to exploiting various weakly supervised methods which leverage numerous unlabeled data. However, many aspects with regard to some unique properties of AUs, such as the regional and relational characteristics, are not sufficiently explored in previous works. Motivated by this, we take the AU properties into consideration and propose two auxiliary AU related tasks to bridge the gap between limited annotations and the model performance in a self-supervised manner via the unlabeled data. Specifically, to enhance the discrimination of regional features with AU relation embedding, we design a task of RoI inpainting to recover the randomly cropped AU patches. Meanwhile, a single image based optical flow estimation task is proposed to leverage the dynamic change of facial muscles and encode the motion information into the global feature representation. Based on these two self-supervised auxiliary tasks, local features, mutual relation and motion cues of AUs are better captured in the backbone network. Furthermore, by incorporating semi-supervised learning, we propose an end-to-end trainable framework named weakly supervised regional and temporal learning (WSRTL) for AU recognition. Extensive experiments on BP4D and DISFA demonstrate the superiority of our method and new state-of-the-art performances are achieved.

cs.CV↗

arXiv2022

Learning Multiple Explainable and Generalizable Cues for Face Anti-spoofing

Although previous CNN based face anti-spoofing methods have achieved promising performance under intra-dataset testing, they suffer from poor generalization under cross-dataset testing. The main reason is that they learn the network with only binary supervision, which may learn arbitrary cues overfitting on the training dataset. To make the learned feature explainable and more generalizable, some researchers introduce facial depth and reflection map as the auxiliary supervision. However, many other generalizable cues are unexplored for face anti-spoofing, which limits their performance under cross-dataset testing. To this end, we propose a novel framework to learn multiple explainable and generalizable cues (MEGC) for face anti-spoofing. Specifically, inspired by the process of human decision, four mainly used cues by humans are introduced as auxiliary supervision including the boundary of spoof medium, moiré pattern, reflection artifacts and facial depth in addition to the binary supervision. To avoid extra labelling cost, corresponding synthetic methods are proposed to generate these auxiliary supervision maps. Extensive experiments on public datasets validate the effectiveness of these cues, and state-of-the-art performances are achieved by our proposed method.

cs.CV↗

arXiv2021

Age of Information in Energy Harvesting Aided Massive Multiple Access Networks

Given the proliferation of the massive machine type communication devices (MTCDs) in beyond 5G (B5G) wireless networks, energy harvesting (EH) aided next generation multiple access (NGMA) systems have drawn substantial attention in the context of energy-efficient data sensing and transmission. However, without adaptive time slot (TS) and power allocation schemes, NGMA systems relying on stochastic sampling instants might lead to tardy actions associated both with high age of information (AoI) as well as high power consumption. For mitigating the energy consumption, we exploit a pair of sleep-scheduling policies, namely the multiple vacation (MV) policy and start-up threshold (ST) policy, which are characterized in the context of three typical multiple access protocols, including time-division multiple access (TDMA), frequency-division multiple access (FDMA) and non-orthogonal multiple access (NOMA). Furthermore, we derive closed-form expressions for the MTCD system's peak AoI, which are formulated as the optimization objective under the constraints of EH power, status update rate and stability conditions. An exact linear search based algorithm is proposed for finding the optimal solution by fixing the status update rate. As a design alternative, a low complexity concave-convex procedure (CCP) is also formulated for finding a near-optimal solution relying on the original problem's transformation into a form represented by the difference of two convex problems. Our simulation results show that the proposed algorithms are beneficial in terms of yielding a lower peak AoI at a low power consumption in the context of the multiple access protocols considered.

cs.IT↗

arXiv2021

Self-Supervised Regional and Temporal Auxiliary Tasks for Facial Action Unit Recognition

Automatic facial action unit (AU) recognition is a challenging task due to the scarcity of manual annotations. To alleviate this problem, a large amount of efforts has been dedicated to exploiting various methods which leverage numerous unlabeled data. However, many aspects with regard to some unique properties of AUs, such as the regional and relational characteristics, are not sufficiently explored in previous works. Motivated by this, we take the AU properties into consideration and propose two auxiliary AU related tasks to bridge the gap between limited annotations and the model performance in a self-supervised manner via the unlabeled data. Specifically, to enhance the discrimination of regional features with AU relation embedding, we design a task of RoI inpainting to recover the randomly cropped AU patches. Meanwhile, a single image based optical flow estimation task is proposed to leverage the dynamic change of facial muscles and encode the motion information into the global feature representation. Based on these two self-supervised auxiliary tasks, local features, mutual relation and motion cues of AUs are better captured in the backbone network with the proposed regional and temporal based auxiliary task learning (RTATL) framework. Extensive experiments on BP4D and DISFA demonstrate the superiority of our method and new state-of-the-art performances are achieved.

cs.CV↗

arXiv2021

Self-Domain Adaptation for Face Anti-Spoofing

Although current face anti-spoofing methods achieve promising results under intra-dataset testing, they suffer from poor generalization to unseen attacks. Most existing works adopt domain adaptation (DA) or domain generalization (DG) techniques to address this problem. However, the target domain is often unknown during training which limits the utilization of DA methods. DG methods can conquer this by learning domain invariant features without seeing any target data. However, they fail in utilizing the information of target data. In this paper, we propose a self-domain adaptation framework to leverage the unlabeled test domain data at inference. Specifically, a domain adaptor is designed to adapt the model for test domain. In order to learn a better adaptor, a meta-learning based adaptor learning algorithm is proposed using the data of multiple source domains at the training step. At test time, the adaptor is updated using only the test domain data according to the proposed unsupervised adaptor loss to further improve the performance. Extensive experiments on four public datasets validate the effectiveness of the proposed method.

cs.CV↗

arXiv2021

Multi-Level Adaptive Region of Interest and Graph Learning for Facial Action Unit Recognition

In facial action unit (AU) recognition tasks, regional feature learning and AU relation modeling are two effective aspects which are worth exploring. However, the limited representation capacity of regional features makes it difficult for relation models to embed AU relationship knowledge. In this paper, we propose a novel multi-level adaptive ROI and graph learning (MARGL) framework to tackle this problem. Specifically, an adaptive ROI learning module is designed to automatically adjust the location and size of the predefined AU regions. Meanwhile, besides relationship between AUs, there exists strong relevance between regional features across multiple levels of the backbone network as level-wise features focus on different aspects of representation. In order to incorporate the intra-level AU relation and inter-level AU regional relevance simultaneously, a multi-level AU relation graph is constructed and graph convolution is performed to further enhance AU regional features of each level. Experiments on BP4D and DISFA demonstrate the proposed MARGL significantly outperforms the previous state-of-the-art methods.

cs.CV↗

arXiv2020

Mobile Data Transactions in Device-to-Device Communication Networks: Pricing and Auction

Device-to-Device (D2D) communication is offering smart phone users a choice to share files with each other without communicating with the cellular network. In this paper, we discuss the behaviors of two characters in the D2D data transaction model from an economic point of view: the data buyers who wish to buy a certain quantity of data, as well as the data sellers who wish to sell data through the D2D network. The optimal price and purchasing strategies are analyzed and deduced based on game theory.

cs.NI↗

arXiv2020

Aggressive Congestion Control Mechanism for Space Systems

How to implement an impeccable space system-of-systems (SoS) internetworking architecture has been a significant issue in system engineering for years. Reliable data transmission is considered one of the most important technologies of space SoS internetworking systems. Due to the high bit error rate (BER), long time delay and asymmetrical channel in the space communication environment, the congestion control mechanism of classic transport control protocols (TCP) shows unsatisfying performances. With the help of existing TCP modifications, this paper contributes an aggressive congestion control mechanism. The proposed mechanism is characterized with a fast start procedure, as well as the feedback information to analyze network traffic and with a link terminating processing mechanism, which can help to reveal the real reason of packet loss, and maintain the size of congestion window at a high level. Simulation results are shown in the end to verify the proposed scheme.

cs.NI↗

arXiv2020

Complex Network Theoretical Analysis on Information Dissemination over Vehicular Networks

How to enhance the communication efficiency and quality on vehicular networks is one critical important issue. While with the larger and larger scale of vehicular networks in dense cities, the real-world datasets show that the vehicular networks essentially belong to the complex network model. Meanwhile, the extensive research on complex networks has shown that the complex network theory can both provide an accurate network illustration model and further make great contributions to the network design, optimization and management. In this paper, we start with analyzing characteristics of a taxi GPS dataset and then establishing the vehicular-to-infrastructure, vehicle-to-vehicle and the hybrid communication model, respectively. Moreover, we propose a clustering algorithm for station selection, a traffic allocation optimization model and an information source selection model based on the communication performances and complex network theory.

cs.NI↗

arXiv2020

Access Strategy in Super WiFi Network Powered by Solar Energy Harvesting: A POMDP Method

The recently announced Super Wi-Fi Network proposal in United States is aiming to enable Internet access in a nation-wide area. As traditional cable-connected power supply system becomes impractical or costly for a wide range wireless network, new infrastructure deployment for Super Wi-Fi is required. The fast developing Energy Harvesting (EH) techniques receive global attentions for their potential of solving the above power supply problem. It is a critical issue, from the user's perspective, how to make efficient network selection and access strategies. Unlike traditional wireless networks, the battery charge state and tendency in EH based networks have to be taken into account when making network selection and access, which has not been well investigated. In this paper, we propose a practical and efficient framework for multiple base stations access strategy in an EH powered Super Wi-Fi network. We consider the access strategy from the user's perspective, who exploits downlink transmission opportunities from one base station. To formulate the problem, we used Partially Observable Markov Decision Process (POMDP) to model users' observations on the base stations' battery situation and decisions on the base station selection and access. Simulation results show that our methods are efficacious and significantly outperform the traditional widely used CSMA method.

cs.IT↗

arXiv2020

Thirty Years of Machine Learning: The Road to Pareto-Optimal Wireless Networks

Future wireless networks have a substantial potential in terms of supporting a broad range of complex compelling applications both in military and civilian fields, where the users are able to enjoy high-rate, low-latency, low-cost and reliable information services. Achieving this ambitious goal requires new radio techniques for adaptive learning and intelligent decision making because of the complex heterogeneous nature of the network structures and wireless services. Machine learning (ML) algorithms have great success in supporting big data analytics, efficient parameter estimation and interactive decision making. Hence, in this article, we review the thirty-year history of ML by elaborating on supervised learning, unsupervised learning, reinforcement learning and deep learning. Furthermore, we investigate their employment in the compelling applications of wireless networks, including heterogeneous networks (HetNets), cognitive radios (CR), Internet of things (IoT), machine to machine networks (M2M), and so on. This article aims for assisting the readers in clarifying the motivation and methodology of the various ML algorithms, so as to invoke them for hitherto unexplored services as well as scenarios of future wireless networks.

cs.NI↗

arXiv2020

Efficient Sampling Algorithms for Approximate Temporal Motif Counting (Extended Version)

A great variety of complex systems ranging from user interactions in communication networks to transactions in financial markets can be modeled as temporal graphs, which consist of a set of vertices and a series of timestamped and directed edges. Temporal motifs in temporal graphs are generalized from subgraph patterns in static graphs which take into account edge orderings and durations in addition to structures. Counting the number of occurrences of temporal motifs is a fundamental problem for temporal network analysis. However, existing methods either cannot support temporal motifs or suffer from performance issues. In this paper, we focus on approximate temporal motif counting via random sampling. We first propose a generic edge sampling (ES) algorithm for estimating the number of instances of any temporal motif. Furthermore, we devise an improved EWS algorithm that hybridizes edge sampling with wedge sampling for counting temporal motifs with 3 vertices and 3 edges. We provide comprehensive analyses of the theoretical bounds and complexities of our proposed algorithms. Finally, we conduct extensive experiments on several real-world datasets, and the results show that our ES and EWS algorithms have higher efficiency, better accuracy, and greater scalability than the state-of-the-art sampling method for temporal motif counting.

cs.SI↗

arXiv2019

Human-Like Decision Making: Document-level Aspect Sentiment Classification via Hierarchical Reinforcement Learning

Recently, neural networks have shown promising results on Document-level Aspect Sentiment Classification (DASC). However, these approaches often offer little transparency w.r.t. their inner working mechanisms and lack interpretability. In this paper, to simulating the steps of analyzing aspect sentiment in a document by human beings, we propose a new Hierarchical Reinforcement Learning (HRL) approach to DASC. This approach incorporates clause selection and word selection strategies to tackle the data noise problem in the task of DASC. First, a high-level policy is proposed to select aspect-relevant clauses and discard noisy clauses. Then, a low-level policy is proposed to select sentiment-relevant words and discard noisy words inside the selected clauses. Finally, a sentiment rating predictor is designed to provide reward signals to guide both clause and word selection. Experimental results demonstrate the impressive effectiveness of the proposed approach to DASC over the state-of-the-art baselines.

cs.CL↗

arXiv2019

On Deflection of Solar Coronal Mass Ejections by the Ambient Coronal Magnetic Field Configuration

Solar Coronal Mass Ejections (CMEs) are sometimes deflected during their propagation. This deflection may be the consequence of interaction between a CME and a coronal hole or the solar wind. We analyze 44 halo-CMEs whose deflection angle exceeds 90 degrees. The coronal magnetic field configuration is computed from daily synoptic maps of magnetic field from SOHO/MDI and SDO/HMI using a Potential Field Source Surface (PFSS) model. By comparing the ambient magnetic field configuration and the measured position angles (MPA) of the CMEs, we conclude that the deflection of 80% of the CMEs (35 of 44) are consistent with the ambient magnetic field configuration, agreeing with previous studies. Of these 35, 71% are deflected toward the heliospheric current sheet (HCS), and 29 degrees toward a pseudo-streamer (PS), the boundary between the same-polarity magnetic field regions. This implies that the ambient coronal magnetic field configuration plays an important and major role in the deflection of CMEs, and that the HCS configuration is more important than PS. If we exclude 13 CMEs having much higher uncertainty from the sample, the agreement between the deflection of CMEs and the ambient field configuration increases substantially, reaching 94% in the new sample of 31 CMEs.

astro-ph.SR↗

arXiv2019

Aeronautical Ad Hoc Networking for the Internet-Above-The-Clouds

The engineering vision of relying on the ``smart sky" for supporting air traffic and the ``Internet above the clouds" for in-flight entertainment has become imperative for the future aircraft industry. Aeronautical ad hoc Networking (AANET) constitutes a compelling concept for providing broadband communications above clouds by extending the coverage of Air-to-Ground (A2G) networks to oceanic and remote airspace via autonomous and self-configured wireless networking amongst commercial passenger airplanes. The AANET concept may be viewed as a new member of the family of Mobile ad hoc Networks (MANETs) in action above the clouds. However, AANETs have more dynamic topologies, larger and more variable geographical network size, stricter security requirements and more hostile transmission conditions. These specific characteristics lead to more grave challenges in aircraft mobility modeling, aeronautical channel modeling and interference mitigation as well as in network scheduling and routing. This paper provides an overview of AANET solutions by characterizing the associated scenarios, requirements and challenges. Explicitly, the research addressing the key techniques of AANETs, such as their mobility models, network scheduling and routing, security and interference are reviewed. Furthermore, we also identify the remaining challenges associated with developing AANETs and present their prospective solutions as well as open issues. The design framework of AANETs and the key technical issues are investigated along with some recent research results. Furthermore, a range of performance metrics optimized in designing AANETs and a number of representative multi-objective optimization algorithms are outlined.

cs.NI↗