Search

Safety of drivers in freeway work zones has been a problem. Real-time crash prediction helps prevent crashes before they happen. This paper looks at real-time crash prediction in freeway work zones by using machine learning approaches. Both the Convolutional Neural Network and the Binary Logistic Regression model are introduced. For training and testing the models, crash data and traffic data from several freeways in D7 zone, Los Angeles, California, were used. Crash data were collected from California Highway Patrol Incident System, and traffic data were obtained from the Caltrans Performance Measurement System. Data processing and matching were conducted. Both the two models were trained and tested. Results show that the Convolutional Neural Network performed slightly better over the Binary Logistic Regression model in predicting crashes with a global accuracy of 79.50%. Despite this, the main merit of the Binary Logistic Regression model is that it is able estimate the impact of affecting variables on the probability of crashes and can help identify the factors related to risks in work zones. Machine learning approaches applied in this study perform well in crash prediction. In general, machine learning techniques and reliable real-time crash prediction applications can be promising in helping drivers and transportation engineers make timely responses to potential crashes on freeways.

We propose a novel method of exploiting informative video segments by learning segment weights for temporal action localization in untrimmed videos. Informative video segments represent the intrinsic motion and appearance of an action, and thus contribute crucially to action localization. The learned segment weights represent the informativeness of video segments to recognizing actions and help infer the boundaries required to temporally localize actions. We build a supervised temporal attention network (STAN) that includes a supervised segment-level attention module to dynamically learn the weights of video segments, and a feature-level attention module to effectively fuse multiple features of segments. Through the cascade of the attention modules, STAN exploits informative video segments and generates descriptive and discriminative video representations. We use a proposal generator and a classifier to estimate the boundaries of actions and classify the classes of actions. Extensive experiments are conducted on two public benchmarks: THUMOS2014 and ActivityNet1.3. The results demonstrate that our proposed method achieves competitive performance compared with the state-of-the-art methods. Moreover, compared with the baseline method that equally treats video segments, STAN achieves significant improvements with the mAP increased from 30.4% to 39.8% on the THUMOS2014 dataset and from 31.4% to 35.9% on the ActivityNet1.3 dataset, demonstrating the effectiveness of learning informative video segments for temporal action localization.

Internet of things (IoT) has attracted considerable attention in recent years due to its potential of interconnecting a large number of heterogeneous wireless devices. However, it is usually challenging to provide reliable and efficient random access control when massive IoT devices are trying to access the network simultaneously. In this paper, we investigate methods to introduce intelligent random access management for a massive cellular IoT network to reduce access latency and access failures. Towards this end, we introduce two novel frameworks, namely local device selection (LDS) and intelligent preamble selection (IPS). LDS enables local communication between neighboring devices to provide cluster-wide cooperative congestion control, which leads to a better distribution of the access intensity under bursty traffics. Taking advantage of the capability of reinforcement learning in developing cooperative multi-agent policies, IPS is introduced to enable the optimization of the preamble selection policy in each IoT clusters. To handle the exponentially growing action space in IPS, we design a novel reinforcement learning structure, named branching actor-critic, to ensure that the output size of the underlying neural networks only grows linearly with the number of action dimensions. Simulation results indicate that the introduced mechanism achieves much lower access delays with fewer access failures in various realistic scenarios of interests.

When an orbital-angular-momentum (OAM) beam propagates through the dynamic air–water interface, the aerosol above the water and the water surface curvature could induce various degradations (e.g., wavefront distortion, beam wandering, scattering, and absorption). Such time-varying degradations could affect the received intensity and phase profiles of the OAM beams, resulting in dynamic modal power loss and modal power coupling. We experimentally investigate the degradation for a single OAM beam under dynamic aerosol, dynamic curvature, and their comprehensive effects. Our results show the following: (i) with the increase of the aerosol strength (characterized by the attenuation coefficient) from ∼0 to ∼0.7–1.3 dB/cm over ∼7 cm, the power coupling ratio from OAM −1 to +2 increases by 4 dB, which might be due to the amplitude and phase distortion caused by spatially dependent scattering and absorption. (ii) With the increase of the curvature strength (characterized by the variance of curvature slope over time) from ∼0 to ∼2 × 10−5 rad2, the power coupling ratio from OAM −1 to +2 increases by 11 dB. This could be caused by both the wavefront distortion and the beam wandering. (iii) Under the comprehensive effect of aerosol (∼0.1–0.6 dB/cm) and curvature (∼6 × 10−7 rad2), there is an up to 2 dB higher modal power loss as compared with the single-effect cases. (iv) The received power on OAM −1 fluctuates in a range of ∼6 dB within a 220 ms measurement time under aerosol (∼0.1–0.6 dB/cm) and curvature (∼6 × 10−7 rad2) effects due to the dynamic degradations. We also demonstrate an OAM −1 and +2 multiplexed 2-Gbit/s on–off-keying link under dynamic aerosol and curvature effects. The results show a power penalty of ∼3 dB for the bit-error-rate at the 7% forward-error-correction limit under the comprehensive effect of aerosol (∼0.1–0.6 dB/cm) and curvature (∼6 × 10−7 rad2), compared with the no-effect case.

In order to make the superheated steam flowing at high speed complete the desuperheating process in the pipeline with limited distance, so as to save the production cost, this paper makes structural innovation on the existing equipment, a solution to install static mixing elements in the pipeline downstream of the nozzle of the desuperheater is proposed to improve the performance of the desuperheater without increasing the cost of nozzle atomization. Based on Fluent software, the working process of the new desuperheater is numerically simulated, and its working mechanism and desuperheating effect are visually analyzed. Compared with the traditional desuperheater, the desuperheating range of the new desuperheater in the pipeline with the same distance is increased by about 70.8%, and the temperature non-uniformity coefficient is reduced by about 75.5%. Furthermore, the validity of the above simulation results is verified by experiments.

Wireless sensor networks (WSNs) are extensively deployed to collect various data. Due to harsh environments and limitation of computing and communication capabilities of sensor nodes, the quality and reliability of sensor data are compromised by outliers. With the advent of 5G, sensors tend to generate increasingly more complex data. When faced with big data, traditional outlier detection methods relied on sensor nodes and remote cloud are unable to accord satisfactory performance in terms of delay and energy consumption. To address this problem, we propose a mobile edge–cloud collaboration outlier detection framework. Outlier detection is performed by edge nodes between the remote cloud and the underlying WSNs, while the training and updating of detection model are conducted on the cloud. A fast angle‐based outlier detection method is developed to obtain training data. The detection model is constructed based on support vector data description. An on‐line learning‐based iterative optimization scheme is devised to update the detection model. Besides, a fuzzy concept is incorporated into the detection model to alleviate the problem of loose decision boundary. Extensive experiments are conducted on real‐world data set. Simulation results show that our model is superior to three popular methods in terms of delay and energy consumption. In addition, when the percentage of operational nodes is 60%, our proposal prolongs the network lifetime by 14.2% to 69.8% compared to the three methods.

As one of the largest and most representative families of natural medicines harvested from plants, the mass production of opioids legitimately occupies large, worldwide farmland cultivation of opiu...

In random network coding (RNC)-enabled networks, a sufficient number of linearly independent encoded packets, i.e., generations, is guaranteed at the cost of high latency and energy consumption at the transmitter. This letter aims to reduce the weighted sum of latency and energy while guaranteeing a targeted successful transmission probability, which is defined as the probability that the RNC-enabled network has sufficient generations when the transmitter receives a required number of acknowledgements (ACKs) feedback. An analytical framework and a deep learning-based method are developed. Results show that the latency and energy consumption are reduced by jointly optimizing the required number of ACKs and the transmit power at the transmitter.

Background Mesenchymal stem cells (MSCs) are a promising treatment for acute rejection (AR) after heart transplantation (HTx) owing to their immunomodulatory functions by promoting the transformation of macrophages from the M0 to M2 phenotype. However, it is undetermined whether surface expression of C-X-C motif chemokine receptor 4 (CXCR4) by MSCs influences macrophage polarization. In this study, we investigated the effects of MSCs on macrophages caused by CXCR4, and detected the underlying mechanism, which may contribute to improving HTx outcomes. Methods The MSCs were extracted from rat bone marrow and identified using flow cytometry. We subsequently observed the effects of CXCR4 and anti-miRNA-204-3p on cell proliferation and migration, and the effects on macrophage polarization. Dual luciferase reporter assay was used to explore whether miRNA-204-3p was an upstream microRNA (miRNA) of CXCR4. A series of rescue experiments were performed to further confirm the inhibitory effect of miRNA-204-3p on CXCR4. Results The results showed that CXCR4 could promote the proliferation and migration of MSCs. Furthermore, it facilitated MSC-mediated macrophage transformation from the M0 to M2 phenotype. In addition, miRNA-204-3p inhibited the function of CXCR4 of MSCs. Conclusions Regulated by miRNA-204-3p, CXCR4 could inhibit the progression of AR after HTx. This study provides a new insight of the treatment of AR after HTx.

Dynamic spectrum access (DSA) has the great potential to alleviate spectrum shortage and promote network capacity. However, two fundamental technical issues have to be addressed, namely, interference coordination between DSA users and interference suppression for primary users (PUs). These two issues are very challenging since generally there is no powerful infrastructures in DSA networks to support centralized control. As a result, DSA users have to perform spectrum management individually, including spectrum access and power allocation, without accurate channel state information and centralized control. In this article, a novel spectrum management framework is proposed, in which $Q$ -learning, a type of reinforcement learning, is utilized to enable DSA users to carry out effective spectrum management individually and intelligently. For more efficient process, neural networks (NNs) are employed to implement $Q$ -learning processes, so-called deep $Q$ -network (DQN). Furthermore, we also investigate the optimal way to construct DQN considering both the performance of wireless communications and the difficulty of NN training. Finally, extensive simulation studies are conducted to demonstrate the effectiveness of the proposed spectrum management framework.

A time-dependent change in the refractive index of a material leads to a change in the frequency of an optical beam passing through that medium. Here, we experimentally demonstrate that this effect-known as adiabatic frequency conversion (AFC)-can be significantly enhanced by a nonlinear epsilon-near-zero-based (ENZ-based) plasmonic metasurface. Specifically, by using a 63-nm-thick metasurface, we demonstrate a large, tunable, and broadband frequency shift of up to ∼11.2 THz with a pump intensity of 4 GW/cm2. Our results represent a decrease of ∼10 times in device thickness and 120 times in pump peak intensity compared with the cases of bare, thicker ENZ materials for the similar amount of frequency shift. Our findings might potentially provide insights for designing efficient time-varying metasurfaces for the manipulation of ultrafast pulses.

In industrial sensor networks, complex industrial environments may be encountered leading to a mix of signals of different types. Complicated interference caused by mixed signals on industrial equipments may significantly degrade the classification rate of signals, which may result in a long training time in order to extract features. In addition, with limited channel resources, it is difficult to make the global optimal decision in industrial distributed wireless sensor networks. To address this problem, a signal classification method using feature fusion is proposed for industrial Internet of Things in this article. In the proposed method, the received signals of nodes are processed by frequency reduction and sampling pretreatment, based on which intelligent representations of signals are obtained. Using federated learning, the data samples are trained with the feature fusion network. Moreover, the trained deep learning network is used on each sensor node to classify signals, the results of which will be transmitted to aggregation center. In the aggregation center, the improved evidence theory method is used to aggregate the recognition results of each sensor node to achieve the final classification. Simulation shows that the proposed method has excellent classification performances. Notably, it is not required for the proposed method to transmit signals from nodes to the aggregation center, which could effectively protect the privacy of industrial information.

BACKGROUND Gastric intestinal metaplasia (GIM) is a precancerous lesion of the stomach, which severely affects human life and health. Currently, a variety of endoscopic techniques are used to screen/evaluate GIM. Traditional white-light endoscopy (WLE) and acetic-acid chromoendoscopy combined with magnifying endoscopy (ME-AAC) are the interventions of choice due to their diagnostic efficacy for GIM. Optical-enhanced magnifying endoscopy (ME-OE) is a new virtual chromoendoscopy technique to identify GIM, which combines bandwidth-limited light and image enhancement processing technology to enhance the detection of mucosal and vascular details. We hypothesized that ME-OE is superior to WLE and ME-AAC in the evaluation of GIM. AIM To directly compare the diagnostic value of WLE, ME-AAC, and ME-OE for detection of GIM. METHODS A total of 156 patients were subjected to consecutive upper gastrointestinal endoscopy examinations using WLE, ME-AAC, and ME-OE. Histopathological findings were utilized as the reference standard. Accuracy, sensitivity, specificity, and positive and negative predictive values of the three endoscopy methods in the diagnosis of GIM were evaluated. Moreover, the time to diagnosis with ME-AAC and ME-OE was analyzed. Two experts and two non-experts evaluated the GIM images diagnosed using ME-OE, and diagnostic accuracy and intra- and inter-observer agreement were analyzed. RESULTS GIM was detected in 68 of 156 patients (43.6%). The accuracy of ME-OE was highest (91.7%), followed by ME-AAC (86.5%), while that of WLE (51.9%) was lowest. Per-site analysis showed that the overall diagnostic accuracy of ME-OE was higher than that of ME-AAC (P = 0.011) and WLE (P < 0.001). The average diagnosis time was lower in ME-OE than in ME-AAC (64 ± 7 s vs 151 ± 30 s, P < 0.001). Finally, the inter-observer agreement was strong for both experts (k = 0.862) and non-experts (k = 0.800). The internal consistency was strong for experts (k = 0.713, k = 0.724) and moderate for non-experts (k = 0.667, k = 0.598). CONCLUSION For endoscopists, especially experienced endoscopists, ME-OE is an efficient, convenient, and time-saving endoscopic technique that should be used for the diagnosis of GIM.

With the proliferation of mobile access services and wireless devices, spectrum resources are increasingly becoming scarce. Industrial wireless sensor networks may have to share frequency bands with other systems and suffer from considerable interference. To address that, industrial cognitive radio networks (ICRNs) were developed for effective spectrum sharing, where signal classification is a fundamental and important technology, especially for industrial wireless devices, which need to identify suspicious transmissions. In this article, a novel framework of signal intelligent classification is proposed based on deep learning networks in ICRNs. In the proposed framework, wireless signals will be preprocessed first by Choi–Williams distribution time–frequency analysis and represented by two-dimensional time–frequency images. Then, features of wireless signals are extracted through stack hybrid autoencoders (SHAEs). To accommodate general cases, we design multiple signal classification methods, including unsupervised, semisupervised, and supervised methods, which are processed by Softmax function, semisupervised linear discriminant function, and Fisher discriminant function, respectively. Finally, simulation studies are conducted and the corresponding simulation results show that the proposed framework is able to learn hierarchical features accurately and achieve excellent signal classification performance. Moreover, it can effectively overcome the negative impact caused by feature parameters uncertainty.

A light beam’s frequency can blueshift when the beam travels through a medium that exhibits a time-dependent decrease in the refractive index. Here we show that a metasurface made of a plasmonic an...

Abnormal event detection aims to identify the events that deviate from expected normal patterns. Existing methods usually extract normal spatio-temporal patterns of appearance and motion in a separate manner, which ignores low-level correlations between appearance and motion patterns and may fall short of capturing fine-grained spatio-temporal patterns. In this paper, we propose to simultaneously learn appearance and motion to obtain fine-grained spatio-temporal patterns. To this end, we present an adversarial 3D convolutional auto-encoder to learn the normal spatio-temporal patterns and then identify abnormal events by diverging them from the learned normal patterns in videos. The encoder captures the low-level correlations between spatial and temporal dimensions of videos, and generates distinctive features representing visual spatio-temporal information. The decoder reconstrucccts the original video from the encoded features representing by 3D de-convolutions and learns the normal spatio-temporal patterns in an unsupervised manner. We introduce the denoising reconstruction error and adversarial learning strategy to train the 3D convolutional auto-encoder to implicitly learn accurate data distributions that are considered normal patterns, which benefits enhancing the reconstruction ability of the auto-encoder to discriminate abnormal events. Both the theoretical analysis and the extensive experiments on four publicly available datasets demonstrate the effectiveness of our method.

Orbital-angular-momentum (OAM) multiplexing has been utilized to increase the channel capacity in both millimeter-wave and optical domains. Terahertz (THz) wireless communication is attracting increasing attention due to its broadband spectral resources. Thus, it might be valuable to explore the system performance of THz OAM links to further increase the channel capacity. In this paper, we study through simulations the fundamental system-degrading effects when using multiple OAM beams in THz communications links under atmospheric turbulence. We simulate and analyze the effects of divergence, turbulence, limited-size aperture, and misalignment on the signal power and crosstalk of THz OAM links. We find through simulations that the system-degrading effects are different in two scenarios with atmosphere turbulence: (a) when we consider the same strength of phasefront distortion, faster divergence (i.e., lower frequency; smaller beam waist) leads to higher power leakage from the transmitted mode to neighbouring modes; and (b) however, when we consider the same atmospheric turbulence, the divergence effect tends to affect the power leakage much less, and the power leakage increases as the frequency, beam waist, or OAM order increases. Simulation results show that: (i) the crosstalk to the neighbouring mode remains

Aspect-based sentiment analysis (ABSA) aims to analyze the sentiment polarity of an input sentence in a certain aspect. Many existing methods of ABSA employ long short-term memory (LSTM) networks and attention mechanism. However, the attention mechanism only models the local certain dependencies of the input information, which fails to capture the global dependence of the inputs. Simply improving the attention mechanism fails to solve the issue of target-sensitive sentiment expression, which has been proven to degrade the prediction effectiveness. In this work, we propose the multi-head self-attention transformation (MSAT) networks for ABSA tasks, which conducts more effective sentiment analysis with target specific self-attention and dynamic target representation. Given a set of review sentences, MSAT applies multi-head target specific self-attention to better capture the global dependence and introduces target-sensitive transformation to effectively tackle the problem of target-sensitive sentiment at first. Second, the part-of-speech (POS) features are integrated into MSAT to capture the grammatical features of sentences. A series of experiments carried on the SemEval 2014 and Twitter datasets show that the proposed model achieves better effectiveness compared with several state-of-the-art methods.

Background Chronic cerebral hypoperfusion (CCH) is a common pathophysiological mechanism in neurodegenerative diseases, such as Alzheimer’s disease and vascular dementia. The orphan nuclear receptor TLX plays an important role in neural development, adult neurogenesis and cognition. The aim of this study was to investigate the neuroprotective effects of TLX on cognitive dysfunction, hippocampal neurogenesis and neuroinflammation in a rat model of CCH and to assess the possible mechanisms. Methods Permanent bilateral common carotid artery occlusion (2-VO) was used to establish a model of CCH. Stereotaxic injection of an adeno-associated virus vector expressing TLX was used to overexpress TLX in the hippocampus. Cognitive function was evaluated by the Morris Water Maze test. Immunofluorescent staining was used to assess hippocampal neurogenesis. The effects of overexpression of TLX on SIRT1 and inflammatory cytokines were analyzed with qRT-PCR and western blot. Result After 2-VO, CCH rats exhibited cognitive impairment and reduction of hippocampal TLX levels. Overexpression of TLX ameliorated cognitive impairments with increasing number of BrdU + cells and BrdU + NeuN + cells in DG. Furthermore, TLX rescued the reduced SIRT1 usually induced by CCH. Additionally, TLX also inhibited the expression of inflammatory cytokines such as NF-κB and IL-1β. Conclusions The present findings suggested that TLX exerted protective effects against cognitive deficits induced by CCH. The possible mechanisms of TLX may be through regulating the SIRT1/NF-κB pathway, promoting hippocampal neurogenesis and inhibiting the neuroinflammatory response.

Unmanned aerial vehicles (UAVs) are becoming important communication infrastructures. One major challenge of communications with UAV networks is the routing protocol design. Due to the inherent characteristics (e.g., dynamic network topology and limited UAV device capabilities), it is difficult to directly apply existing routing protocols that utilize network topology information and routing path explorations. In this article, two novel routing protocols are designed based on random network coding (RNC) for a swarm UAV network, where UAVs operate cooperatively as a swarm, enabling efficient routing process. The first routing protocol utilizes the unique feature of RNC: Original packets can be decoded as long as an UAV accumulates sufficient generations. This property can be used to effectively expedite the underlying routing process. The second routing protocol further improves the efficiency where each forwarding UAV only needs to create a new generation rather than decoding original packets. Accordingly, the duration of each hop can be significantly reduced. Extensive simulations have been conducted to evaluate the performance of the designed routing protocols. The simulation results demonstrate that our designed routing protocols can effectively enhance the performance on both average transmission delay and delay violation probabilities compared to benchmark methods.

Zeolitic imidazolate framework-8 (ZIF-8)/semiconductor nanocomposites were synthesized by a simple method. The nanocomposites were composed of the in situ grown ZIF-8 and supported ZnIn-layered dou...

Automatically detecting anomalies in videos is a challenging problem due to non-deterministic definitions of abnormal events and lack of sufficient training data. To address these issues, we propose an autoencoder coupled with attention model to discover normal patterns in videos via adversarial learning. Abnormal events are detected by diverging them from the normal patterns with the reconstruction error produced by the autoencoder. To this end, we build an end-to-end trainable adversarial attention-based autoencoder network, called Ada-Net, to make the reconstructed frames indistinguishable from original frames. The Ada-Net combines an autoencoder network and a GAN model that is used to benefit enhancing the reconstruction ability of the autoencoder. To further improve the reconstruction performance, we integrate an attention model into the decoder to dynamically select informative parts of encoding features for decoding. The attenion mechanism is helpful to preserving important information for learning intrinsic normal patterns. Evaluations on four challenging datasets, including the Subway, the UCSD Pedestrian, the CUHK Avenue, and the ShanghaiTech datasets, demonstrate the effectiveness of the proposed method.

Abiotic–biological hybrid systems that combine the advantages of abiotic catalysis and biotransformation for the conversion of carbon dioxide (CO2) to value-added chemicals and fuels have emerged as an appealing way to address the global energy and environmental crisis caused by increased CO2 emission. We illustrate the recent progress in this field. Here, we first review the natural CO2 fixation pathways for an in-depth understanding of the biological CO2 transformation strategy and why a sustainable feed of reducing power is important. Second, we review the recent progress in the construction of abiotic–biological hybrid systems for CO2 transformation from two aspects: (i) microbial electrosynthesis systems that utilize electricity to support whole-cell biological CO2 conversion to products of interest and (ii) photosynthetic semiconductor biohybrid systems that integrate semiconductor nanomaterials with CO2-fixing microorganisms to harness solar energy for biological CO2 transformation. Lastly, we discuss potential approaches for further improvement of abiotic–biological hybrid systems.

Scalable video coding (SVC) and video caching are two promising techniques in the 5th generation networks to improve the users’ quality of experience (QoE) in terms of video retrieval. In this paper, we study the video content retrieval in cache-aided device-to-device (D2D) networks, where each video content is coded into multiple layers via SVC. In video caching placement phase, the probabilistic caching placement policy is applied, while in content retrieval phase, the non-orthogonal transmission scheme is utilized. Besides, different D2D transmitter selection algorithms are considered and cache-aided data rate (CADR) is formulated as a metric in this paper, and it is maximized by jointly optimizing the probability caching policy and the power allocation policy in two phases. Analytical results show that probabilistic caching policy incorporated with power domain non-orthogonal transmission scheme can achieve significant benefits compared with other benchmark schemes.

The structural and optical properties of amorphous silicon (a-Si) and Al-dispersed amorphous silicon (a-Si:Al) thin films irradiated by femtosecond (fs) laser at various energy densities are investigated comparatively in this article. It is found that there is an uneven crystallization in both amorphous thin films by means of optical microscopy and laser Raman spectroscopy respectively. The crystallization in each pulse spot area is gradually weakened from the center to the edge along with the energy dispersion of laser irradiation. The laser induced crystallization in a-Si thin films begins early and develops more extensively compared to that in a-Si:Al thin films, and Al nanoparticles inhibit somehow the crystallization of a-Si in a-Si:Al thin films.

We experimentally generate an orbital-angular-momentum (OAM) beam with a tunable mode order over a range of wavelengths utilizing an integrated broadband pixel-array OAM emitter. The emitter is composed of a 3-to-4 coupler, four phase controllers, and a mode convertor. An optical input is split into four waveguides by the coupler. Subsequently, the four waveguide fields are coherently combined and transformed into a free-space OAM beam by the mode convertor. By tuning the phase delay Δ φ between the four waveguides using the integrated phase controllers, the OAM order of the generated beam could be changed. Our results show that (a) a single OAM beam with a tunable OAM order ( ℓ = − 1 or ℓ = + 1 ) is generated with the intermodal power coupling of < − 11 d B , and (b) in a wavelength range of 6.4 nm, a free-space link of a single 50 Gbaud quadrature-phase-shift-keying (QPSK) channel carried by the tunable OAM beam is achieved with a bit error rate below the forward-error-correction threshold. As proof of concept, a 400 Gbit/s OAM-multiplexed and WDM QPSK link is demonstrated with a ∼ 1 − dB OSNR penalty compared with a single-beam link.

Annular syphilis may range from mildly raised lesions with scaly borders to verrucous plaques. Localized annular syphilis on the genitalia has been rarely reported in HIV-negative cases. This paper reports a case of annular secondary syphilis on the penis. Dermoscopy showed peripheral dotted and short linear vessels and white scaling with a relatively clear central area in an erythematous annular plaque. Histopathology revealed mild hyperkeratosis, parakeratosis, psoriasiform acanthosis, and focal basal vacuolar degeneration with lichenoid, perivascular, and periadnexal infiltrate of lymphohistiocytes and plasma cells in the superficial dermis. Silver stain showed several spirochetes in the lower epidermis and superficial dermis. Electron microscopy revealed a few intercellular and intracytoplasmic spirochetes in the basal epidermis and free spirochetes in the papillary dermis. Rapid plasma reagin and Treponema pallidum particle agglutination assays were positive. The lesions disappeared after intramuscular benzathine penicillin, with no relapse at six-month follow-up.

The explosive growth of wireless devices motivates the development of the internet-of-things (IoT), which is capable of interconnecting massive and diverse "things" via wireless communications. This is also called massive machine type communications (mMTC) as a part of the undergoing fifth generation (5G) mobile networks. It is envisioned that more sophisticated devices would be connected to form a hyperconnected world with the aids of the sixth generation (6G) mobile networks. To enable wireless accesses of such IoT networks, artificial intelligence (AI) can play an important role. In this article, the frameworks of centralized and distributed AI-enabled IoT networks are introduced. Key technical challenges, including random access and spectrum sharing (spectrum access and spectrum sensing), are analyzed for different network architectures. Deep reinforcement learning (DRL)-based strategies are introduced and neural networks-based approaches are utilized to efficiently realize the DRL strategies for system procedures such as spectrum access and spectrum sensing. Different types of neural networks that could be used in IoT networks to conduct DRL are also discussed.

Due to the lack of professional testing and evaluation system for cloud computing information security products, the basic security of cloud computing information security products cannot be guaranteed. The establishment of a test and evaluation system of cloud computing information security products is provided, and the system is used in the actual product testing, to further promote the development of cloud computing and information security.

Semiconductor-based durable photocatalysts for efficient water splitting have attracted much attention for the development of sustainable hydrogen energy production, but it is challenging for CdS to achieve the expectation of the absence of hole scavengers. Herein, it is reported that uniform CdS nanorods coated with ultrathin NiOOH were prepared as a photocatalyst for high-efficiency photocatalytic water splitting without using any hole sacrificial agents. For the novel surface dynamics features, electron accumulation on CdS was detected with in situ irradiated X-ray photoelectron spectroscopy and accelerated hole transfer was recorded as 49.6 ± 9.2 ps by femtosecond time-resolved transient absorption spectroscopy. The thickness dependence of the NiOOH wrapper with an ultimate continuous thickness of ∼4 nm not only achieved the record high value for photocatalytic hydrogen generation rate (118.6 μmol h−1 g−1) among CdS-based heterojunction photocatalysts without any scavengers, but also exhibited good photostability (over 25 h of cycling measurements). This work provides valuable guidelines for the design of next-generation, high performance CdS-based photocatalysts.

In this letter, we experimentally explore the limited-size aperture effects in an orbital-angular-momentum (OAM)-multiplexed free-space optical (FSO) data link for aerial platforms. 200-Gbit/s data transmission is demonstrated between a ground transmitter and a ground receiver via a flying, retro-reflecting unmanned-aerial-vehicle (UAV) over 100-m round-trip distance by multiplexing two OAM beams. Our results indicate that when the receiver aperture size is limited, both power loss and channel crosstalk would increase. With a receiver aperture diameter of >5 cm, bit error rate (BER) mostly below 3.8 × 10−3 has been achieved for both channels

Chinese biochemical engineering is committed to supporting the chemical and food industries, to advance science and technology frontiers, and to meet major demands of Chinese society and national economic development. This paper reviews the development of biochemical engineering, strategic deployment of these technologies by the government, industrial demand, research progress, and breakthroughs in key technologies in China. Furthermore, the outlook for future developments in biochemical engineering in China is also discussed.

In this paper, we present a novel framework of automatically localizing action instances based on action pattern trees (AP-Trees) in a long untrimmed video. For localizing action instances in videos with varied temporal lengths, we first split videos into sequential segments and then use the AP-Trees to produce precise temporal boundaries of action instances. The AP-Trees can exploit the temporal information between segments of videos based on the label vectors of segments, by learning the occurrence frequency and order of segments. In AP-Trees, nodes stand for action class labels of segments and edges represent the temporal relationships between two consecutive segments. Thus, we can discover the occurrence frequencies of segments by searching paths of AP-Trees. In order to obtain accurate labels of video segments, we introduce deep neural networks to annotate the segments by simultaneously leveraging the spatio-temporal information and the high-level semantic feature of segments. In the networks, informative action maps are generated by a global average pooling layer to retain the spatio-temporal information of segments. An overlap loss function is employed to further improve the precision of label vectors of segments by considering the temporal overlap between segments and the ground truth. The experiments on THUMOS2014, MSR ActionII, and MPII Cooking datasets demonstrate the effectiveness of the method.

We use two different methods to fabricate nanostructured silicon on the surface of C-Si: femtosecond laser etching (FLE) and deep reactive ion etching (DRIE) combined with plasma immersion ion implantation (PIII). Nanocone silicon arrays of dense and random distribution are obtained by FLE. Meanwhile, cylindroid silicon nanostructures of excellent regularity and uniform coverage are achieved by DRIE. These nanostructured silicon materials show a remarkable enhancement on absorptance at near-infrared wavelength. Moreover, the minority carriers lifetime measurement is also carried out to evaluate defect states caused by two etching processes and their influence on semiconductor physical effects. A Si-PIN photoelectronic detector with nanostructured silicon at the back surface exhibits high near-infrared responsivity. These novel results may have a potential application in near-infrared photoelectronic devices.

This paper describes a new design of high-frequency differential line for Ex/Hz probe calibration. Both simulation and measurement results show that it can generate a pure transverse electromagnetic field up to 30GHz.

A riser is a fluid conduit connecting from a subsea well to a surface floating production vessel such as a spar, TLP, or a semi-submersible. It is a key component for transporting produced oil and gas from the subsea wells to the surface production vessel. Through nearly 30 years of design and implementation, Steel Catenary Risers (SCRs) have been found to have the advantages of relatively low manufacturing cost and good adaptability to floating platform’s motion. Thus, SCRs are widely used in GoM and Brazilian water. This paper investigates design challenges of SCRs used for a semi-submersible in deepwater and ultra deepwater. As the application moves to deepwater and ultra deepwater, challenges of a conventional SCRs increase significantly in terms of global performance, fabrication, and installation. This paper first introduces an SCR system used for a semi-submersible platform. One of the challenges is the harsh environmental condition such as hurricanes or tropical cyclones. Severe motions of a semi-submersible platform may cause significant compression loads on SCRs at the Touch Down Zone (TDZ). It is challenging to design the SCRs to meet API RP 2RD requirement for the riser dynamic strength. Multiple seeds of random motion simulations for irregular wave are necessary for the evaluation. The new riser design code of API STD 2RD can also be used to assess riser strength, especially under 1000-yr survival environmental conditions. It is challenging to design SCRs to meet a design life of 30 years and beyond. The fatigue assessment of the SCRs considers the damage contributions from the semi wave motions, riser Vortex-Induced Vibration (VIV) due to ambient current, and Vortex-Induced Motion (VIM) of a semi-submersible platform. Fatigue contribution from installation is also considered and included in the combined fatigue damage calculation. The long term combined factored fatigue life is computed to ensure that the SCRs meet the requirement of design life. This paper further investigates intermittent Vortex-Induced Vibration (VIV) caused by heave motions of a semi-submersible platform. This VIV phenomenon has an impact on SCR strength due to the drag amplification as well as fatigue damage. The installation concerns come from riser pipe size, installation method and capacity of an installation vessel. Examples are provided for illustrations of the design challenges. In addition, this paper explores likely solutions to meet SCR design challenges for the applications in deepwater and ultra deepwater.

The J-lay and S-lay are two common methods for SCR and pipeline installations. When using the S-lay installation method, onboard welded pipe joints leave the vessel horizontally and are guided to the seabed over a stinger structure. The pipe is lowered using tensioners. With the advantage of high production rate, Slay can be a cost-effective solution for deepwater riser and pipeline installation. Based on HYSY201 installation vessel, this paper investigates the feasibility of S-lay installation for deepwater SCRs and pipelines to be used in South China Sea. It first introduces the SCRs and pipelines to be used for a deep draft semi-submersible for the Lingshui 17-2 project. It then presents the S-lay installation vessel HYSY201 and S-lay configuration. The hydrodynamic motion analysis for a Response Amplitude Operator (RAO) was computered for HYSY201 in different environmental headings. With the site-specific metocean design basis, this paper presents an installation procedure, analysis methodology, and acceptance criteria. The study covers different sizes of SCRs and pipelines to investigate the feasibility of S-lay installation. The study starts from the static installation analysis of SCRs and pipelines and includes different installation steps. The acceptance criteria are examined for the pipes at over bend and sag bend regions. The support reactions load on the stinger structure are reported at each step. The dynamic analysis is selectively performed to evaluate Dynamic Amplification Factors (DAFs) of support reaction loads especially for roller box supports on the stinger structure. The sensitivity of DAFs to wave parameters such as wave height and peak period is analyzed as well. The extreme support reaction loads are computed for evaluating the strength performance of the stinger structure. The feasibility of S-lay installation for deepwater SCRs and pipelines is determined by the global performance of SCRs and pipelines, installation vessel hold back tension and A&R winch load capacity, and performance of the stinger structure. Based on the study work, this paper finds the feasibility of S-lay installation of deepwater SCRs and pipelines for Lingshui 17-2 project using the installation vessel of HYSY201.

We experimentally investigate the tunable Doppler shift in an 80 nm thick indium-tin-oxide (ITO) film at its epsilon-near-zero (ENZ) region. Under strong and pulsed excitation, ITO exhibits a time-varying change in the refractive index. A maximum frequency redshift of 1.8 THz is observed in the reflected light when the pump light has a peak intensity of ∼ 140 G W / c m 2 and a pulse duration of ∼ 580 f s , at an incident angle of 40°. The frequency shift increases with the increase in pump intensity and saturates at the intensity of ∼ 140 G W / c m 2 . When the pump pulse duration increases from ∼ 580 f s to ∼ 1380 f s , the maximum attainable frequency shift decreases from 1.8 THz to 0.7 THz. In addition, the pump energy required to saturate the frequency shift decreases with the increase in pump pulse duration for ∼ x < 1 p s and remains unchanged for ∼ x > 1 p s durations. Tunability exists among the pump pulse energy, duration, and incident angle for the Doppler shift of the ITO-ENZ material, which can be employed to design efficient frequency shifters for telecom applications.

Traditional modelling and optimization methods for sound absorptive materials require manual testing and huge computation effort. Artificial intelligence is an ideal method for the prediction and identification of complex systems in these cases. This study presents an optimization approach based on image recognition combined with deep learning for cavity resonance sound absorptive material. The sound absorptive material model is conducted using the acoustic-structure coupling model in COMSOL Multiphysics for the sound pressure distribution and behaviour during different frequency excitation. The cavity material plane wave image recognition model was obtained using the CNN method of multiscale layered features. The image arrangement was then processed and sent to the convolutional layer for feature extraction. Subsequently, the feature record was entered into the loop. The association between the kinematic parameters of the cavity material and the sound pressure distribution is obtained through the prediction layer. Using the prediction of CRNN, the results indicate that the range of the cover thickness should be between 50~70 mm and the angle should be between 45~60°. It also showed that this optimization method could be successfully applied for similar materials and components, with certain universality.

Background Hyperthermia has been reported to cause cancer stage regression, thus providing surgical opportunities in patients with unresectable tumors and improving the quality of life of patients by preserving certain organs. Methods A prospective open-label phase II trial was conducted to evaluate the efficacy of hyperthermia combined with induction chemotherapy in patients with locally advanced resectable oral squamous cell carcinoma (OSCC). Patients received hyperthermia combined with two cycles of 5-fluorouracil, cisplatin, and docetaxel (TPF) induction chemotherapy regimens or TPF induction chemotherapy alone, followed by radical surgery with postoperative radiotherapy. The primary endpoint was the clinical response rate of the induction chemotherapy. The secondary endpoints were overall survival (OS), disease-free survival (DFS), and toxicity. Results A total of 120 patients were enrolled, and 115 patients were included in the clinical response analysis. The clinical response rate was significantly higher in the experimental arm than in the control arm (65.45% vs. 40.00%, p = 0.0088). There were no unexpected toxicities, and hyperthermia and induction chemotherapy did not increase the perioperative morbidity rate. Moreover, there was a significant improvement in DFS, but no significant difference in OS between the two arms. In the subgroup analysis, increased OS and DFS rates were associated with patients with favorable clinical response after induction chemotherapy in the total population, experimental arm, and control arm. Conclusions Our study demonstrates that hyperthermia combined with induction chemotherapy is associated with a high response rate and provides a new treatment option for patients with resectable stage III or IVA OSCC.

Post translational modifications (PTM) such as phosphorylation are often correlated with tumorigenesis and malignancy in breast cancer. Herein, we report a PTM-assisted strategy as a simplified version of a personalized cancer vaccine for enhanced cancer immunotherapy. Titanium modified dendritic mesoporous silica nanoparticles (TiDMSN) are applied to assist the specific enrichment of phosphorylated tumor antigens released upon immunogenic cell death. This strategy significantly improved the tumor inhibition efficacy in a bilateral breast cancer model and the expansion of both CD8+ and CD4+ T cells in the distant tumor site. The nanotechnology based PTM-assisted strategy provides a simple and generalizable methodology for effective personalized cancer immunotherapy., The nano-enabled post-translational modification assisted strategy for effective cancer immunotherapy.

We simulate the modal purities and Laguerre-Gaussian (LG) modal coupling of orbital-angular-momentum (OAM) beams reflected by a rough surface for a non-line-of-sight (NLoS) THz communication link. The reflection of the THz OAM beams is simulated using the finite-difference time-domain (FDTD) method. The simulation results show that: (i) an OAM beam reflected by a metal or dielectric rough surface would experience modal coupling in the 2-dimensional LG modal spectra; (ii) reflected by a metal surface, the modal purities of both transverse-electric (TE) and transverse-magnetic (TM) OAM beams increase as the incident angle increases from 0° to 70° or the carrier frequency of the OAM beams decreases from 1 THz to 0.1 THz; and (iii) reflected by a dielectric surface (n = 1.5), the modal purity and received power of the transmitted TE OAM beam also increase with an increase in the incident angle. However, for TM polarization, the modal purity exhibits a dip near the Brewster’s angle. Moreover, with a limited-size receiver aperture, the simulation results for a rough metal surface show that: (i) when the aperture size decreases, the received power of the desired mode decreases by >30 dB while the modal purity of that mode firstly increases and then decreases; and (ii) with the same aperture size, the link could experience ~10 dB higher total power loss and ~12 dB higher channel crosstalk when the standard deviation of the surface height increases from 0.08 mm to 0.20 mm.

Existing routing protocols may not be applicable in UAV networks because of their dynamic network topology and lack of accurate position information. In this paper, an enhanced flooding-based routing protocol is designed based on random network coding (RNC) and clustering for swarm UAV networks, enabling the efficient routing process without any routing path discovery or network topology information. RNC can naturally accelerate the routing process, with which in some hops fewer generations need to be transmitted. To address the issue of numerous hops and further expedite routing process, a clustering method is leveraged, where UAV networks are partitioned into multiple clusters and generations are only flooded from representatives of each cluster rather than flooded from each UAV. By this way, the amount of hops can be significantly reduced. The technical details of the introduced routing protocol are designed. Moreover, to capture the dynamic network topology, the Poisson cluster process is employed to model UAV networks. Afterwards, stochastic geometry tools are utilized to derive the distance distribution between two random selected UAVs and analytically evaluate performance. Extensive simulation studies are conducted to prove the validation of performance analysis, demonstrate the effectiveness of our designed routing protocol, and reveal its design insight.

Background：The clinical characteristics of second primary tumors in oral cavity after radiation was a special subtype of oral cancer and remained poorly studied. The purpose of the present study was to investigate the clinicopathological characteristics and prognostic factors of second primary oral squamous cell carcinoma (OSCC) after radiotherapy for head and neck cancer.Methods：The clinicopathological characteristics of 48 OSCC patients underwent radiotherapy for head and neck cancer were retrospectively analyzed by Kaplan-Meier survival analysis and Cox proportional hazards model, including gender, age, alcohol consumption, smoking, clinical stage, margin status, regional lymph node status, tumor differentiation and treatment mode.Results：The second primary OSCC mostly occurred on the tongue (18/48), bucca (12/48) and gingiva (10/48), and the 3- and 5-year overall survival (OS) was 60.3% and 39.4%, respectively. Margin status (p=0.001, log-rank=10.777) and extranodal extension (p=0.045, log-rank=4.017) were significantly associated with OS, while only margin status was found to be an independent prognostic factor of OS in the Cox proportional hazards model (p=0.003, HR=3.976).Conclusions： OSCC patients underwent radiotherapy for head and neck cancer show poor survival outcomes. Margin status is an independent prognostic factor of second primary OSCC, and the prognostic of second primary OSCC was not in accordance with sporadic OSCC.

The coronavirus disease 19 (COVID-19) pandemic, caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has become the worst public health crisis in a century. However, knowledge about the dynamics of antibody responses in patients with COVID-19 is still poorly understood. In this study, we performed a serological study with serum specimens collected at the acute and the convalescent phases from 104 patients with severe COVID-19 who were part of the first wave of COVID-19 cases in Wuhan, China. Our findings revealed that neutralizing antibodies to SARS-CoV-2 are persistent for at least 6 months in patients with severe COVID-19, despite that IgG levels against the receptor binding domain (RBD) and nucleocapsid protein (N) IgG declined from the acute to the convalescent phase. Moreover, we demonstrate that the level of RBD-IgG is capable of correlating with SARS-CoV-2-neutralizing activities in COVID-19 serum. In summary, our findings identify the magnitude, functionality, and longevity of antibody responses in patients with COVID-19, which sheds light on the humoral immune response to COVID-19 and would be beneficial for developing vaccines.