Search

Yingzi He,&ast; Jingjing Wang,&ast; Jianan Wang, Ruojun Qiu, Shuo Wang, Ting Jin, Hong Li, Fenping Zheng Department of Endocrinology, the Affiliated Sir Run Run Shaw Hospital, College of Medicine, Zhejiang University, Zhejiang, 310016, People’s Republic of China&ast;These authors contributed equally to this workCorrespondence: Fenping Zheng, Department of Endocrinology, the Affiliated Sir Run Run Shaw Hospital, College of Medicine, Zhejiang University, Zhejiang, 310016, People’s Republic of China, Email 3407004@zju.edu.cnObjective: This study assessed possible associations among physical activity (PA), sitting time (ST), metabolic syndrome (MetS), and the individual components thereof. We analyzed the entire study sample and subpopulations stratified by visceral fat area (VFA). We hypothesized that individuals with elevated VFA might respond differently to modifiers of metabolic health, including PA and ST.Methods: This cross-sectional study, conducted between March and May 2010, enrolled 957 adults with abdominal magnetic resonance imaging (MRI) aged 40– 65 years living in the urban communities in Hangzhou, China. PA and ST were recorded using the standard International Physical Activity Questionnaire (IPAQ) and categorized into three levels. The ethnicity-specific cutoff for central obesity was VFA ≥ 80 cm2 on MRI according to Chinese population-based research. Multiple logistic regression models were used to analyze the associations between PA, ST, MetS and its components.Results: In the total subject population, participants reporting high level of PA were at a lower risk of MetS (OR = 0.46, 95% CI: 0.25, 0.86) than those declaring low PA. In the subgroup population with VFA ≥ 80 cm2 (ie, with central obesity), moderate-to-high PA levels were associated with a lower risk of MetS (p for trend < 0.05) and a lower risk of decreased high-density lipoprotein cholesterol (HDL-C) concentrations (p for trend < 0.05). In addition, ST > 3 h

Jingjing Wang,1 Yirui Dong,1 Xuewei Zheng,1 Haodi Ma,1 Mengjiao Huang,1 Dongliao Fu,1 Jiangbo Liu,2 Qinan Yin1,3 1School of Medical Technology and Engineering, Henan University of Science and Technology, Luoyang, People’s Republic of China; 2Department of General Surgery, First Affiliated Hospital, College of Clinical Medicine, Henan University of Science and Technology, Luoyang, People’s Republic of China; 3Henan Engineering Research Center of Digital Pathology and Artificial Intelligence Diagnosis, The First Affiliated Hospital of Henan University of Science and Technology, Luoyang, People’s Republic of ChinaCorrespondence: Qinan Yin, School of Medical Technology and Engineering, Henan University of Science and Technology, No. 263 Kaiyuan Avenue, Luoyang, People’s Republic of China, 471003, Email qinanyin@haust.edu.cn Jiangbo Liu, First Affiliated Hospital, College of Clinical Medicine, Henan University of Science and Technology, Guanlin Avenue, Luoyang, 471031, People’s Republic of China, Email jiangboliuxing@163.comAbstract: Innate immunity is the first line of defense in the human body, and it plays an important role in defending against viral infection. Viruses are identified by different pattern-recognition receptors (PRRs) that activate the mitochondrial antiviral signaling protein (MAVS) or transmembrane protein 173 (STING), which trigger multiple signaling cascades that cause nuclear factor-κB (NF-κB) and interferon regulatory factor 3 (IRF3) to produce inflammatory factors and interferons (IFNs). PRRs play a pivotal role as the first step in pathogen induction of interferon production. Interferon elicits antiviral activity by inducing the transcription of hundreds of IFN-stimulated genes (ISGs) via the janus kinase (JAK) – signal transducer and activator of transcription (STAT) pathway. An increasing number of studies have shown that environmental, pathogen and host factors regulate the IFN signaling pathway. Here, we summarize the mechanism

Cen Cheng, Jingjing Wang, Yuanyuan Cao, Erwei Gu, Xuesheng Liu Department of Anesthesiology, The First Affiliated Hospital of Anhui Medical University, Hefei, People’s Republic of ChinaCorrespondence: Xuesheng Liu, Department of Anesthesiology, The First Affiliated Hospital of Anhui Medical University, Hefei, People’s Republic of China, Email liuxuesheng@ahmu.edu.cnIntroduction: Rectus sheath block is an emerging technique that provide effective perioperative analgesia and is related to lower perioperative opioid consumption and decrease opioid-related adverse effects. The present research is designed to explore the effect of rectus sheath block on recovery quality in patients following transabdominal midline gynecological surgery.Methods: Ninety female patients following elective transabdominal midline gynecological surgery were enrolled. Patients were randomized to group R (n = 45) which receive preoperative ultrasound-guided RSB with 0.4% ropivacaine or group C which is control group (n = 45). The primary outcome was the quality of recovery on the first postoperative day. The quality of recovery was assessed by the 40-item Quality of Recovery questionnaire (QoR-40). Secondary outcomes included the intraoperative opioid consumption, time to first flatus and time to first discharging from bed, postoperative nausea and vomiting, and patient satisfaction.Results: The patients in two groups had comparable baseline characteristics. Postoperative global QoR-40 scores were significantly better in group R than in group C (165.0[159.5-170.0] vs 155.0[150.0-157.0], respectively; median difference 12[95% confidence interval: 8-15, P< 0.001]). Preoperative RSB reduced intraoperative opioid consumption, reduced the time to first flatus, time to first discharging from bed and the post anaesthesia care unit discharge time. Furthermore, group R showed greater patient satisfaction.Conclusion: A single preoperative administration of RSB with ropivacaine improved the quality

Previous face forgery detection methods mainly focus on appearance features, which may be easily attacked by sophisticated manipulation. Considering the majority of current face manipulation methods generate fake faces based on a single frame, which do not take frame consistency and coordination into consideration, artifacts on frame sequences are more effective for face forgery detection. However, current sequence-based face forgery detection methods use general video classification networks directly, which discard the special and discriminative motion information for face manipulation detection. To this end, we propose an effective sequence-based forgery detection framework based on an existing video classification method. To make the motion features more expressive for manipulation detection, we propose an alternative motion consistency block instead of the original motion features module. To make the learned features more generalizable, we propose an auxiliary anomaly detection block. With these two specially designed improvements, we make a general video classification network achieve promising results on three popular face forgery datasets., Comment: Accepted to ICASSP 2023

Gaze estimation methods often experience significant performance degradation when evaluated across different domains, due to the domain gap between the testing and training data. Existing methods try to address this issue using various domain generalization approaches, but with little success because of the limited diversity of gaze datasets, such as appearance, wearable, and image quality. To overcome these limitations, we propose a novel framework called CLIP-Gaze that utilizes a pre-trained vision-language model to leverage its transferable knowledge. Our framework is the first to leverage the vision-and-language cross-modality approach for gaze estimation task. Specifically, we extract gaze-relevant feature by pushing it away from gaze-irrelevant features which can be flexibly constructed via language descriptions. To learn more suitable prompts, we propose a personalized context optimization method for text prompt tuning. Furthermore, we utilize the relationship among gaze samples to refine the distribution of gaze-relevant features, thereby improving the generalization capability of the gaze estimation model. Extensive experiments demonstrate the excellent performance of CLIP-Gaze over existing methods on four cross-domain evaluations., Comment: Accepted to AAAI 2024

Multimodal Conversational Emotion (MCE) detection, generally spanning across the acoustic, vision and language modalities, has attracted increasing interest in the multimedia community. Previous studies predominantly focus on learning contextual information in conversations with only a few considering the topic information in single language modality, while always neglecting the acoustic and vision topic information. On this basis, we propose a model-agnostic Topic-enriched Diffusion (TopicDiff) approach for capturing multimodal topic information in MCE tasks. Particularly, we integrate the diffusion model into neural topic model to alleviate the diversity deficiency problem of neural topic model in capturing topic information. Detailed evaluations demonstrate the significant improvements of TopicDiff over the state-of-the-art MCE baselines, justifying the importance of multimodal topic information to MCE and the effectiveness of TopicDiff in capturing such information. Furthermore, we observe an interesting finding that the topic information in acoustic and vision is more discriminative and robust compared to the language.

Recently, arbitrary-scale point cloud upsampling mechanism became increasingly popular due to its efficiency and convenience for practical applications. To achieve this, most previous approaches formulate it as a problem of surface approximation and employ point-based networks to learn surface representations. However, learning surfaces from sparse point clouds is more challenging, and thus they often suffer from the low-fidelity geometry approximation. To address it, we propose an arbitrary-scale Point cloud Upsampling framework using Voxel-based Network (\textbf{PU-VoxelNet}). Thanks to the completeness and regularity inherited from the voxel representation, voxel-based networks are capable of providing predefined grid space to approximate 3D surface, and an arbitrary number of points can be reconstructed according to the predicted density distribution within each grid cell. However, we investigate the inaccurate grid sampling caused by imprecise density predictions. To address this issue, a density-guided grid resampling method is developed to generate high-fidelity points while effectively avoiding sampling outliers. Further, to improve the fine-grained details, we present an auxiliary training supervision to enforce the latent geometric consistency among local surface patches. Extensive experiments indicate the proposed approach outperforms the state-of-the-art approaches not only in terms of fixed upsampling rates but also for arbitrary-scale upsampling., Comment: Accepted to AAAI 2024. The source code is available at https://github.com/hikvision-research/3DVision

Aspect Sentiment Understanding (ASU) in interactive scenarios (e.g., Question-Answering and Dialogue) has attracted ever-more interest in recent years and achieved important progresses. However, existing studies on interactive ASU largely ignore the coreference issue for opinion targets (i.e., aspects), while this phenomenon is ubiquitous in interactive scenarios especially dialogues, limiting the ASU performance. Recently, large language models (LLMs) shows the powerful ability to integrate various NLP tasks with the chat paradigm. In this way, this paper proposes a new Chat-based Aspect Sentiment Understanding (ChatASU) task, aiming to explore LLMs' ability in understanding aspect sentiments in dialogue scenarios. Particularly, this ChatASU task introduces a sub-task, i.e., Aspect Chain Reasoning (ACR) task, to address the aspect coreference issue. On this basis, we propose a Trusted Self-reflexion Approach (TSA) with ChatGLM as backbone to ChatASU. Specifically, this TSA treats the ACR task as an auxiliary task to boost the performance of the primary ASU task, and further integrates trusted learning into reflexion mechanisms to alleviate the LLMs-intrinsic factual hallucination problem in TSA. Furthermore, a high-quality ChatASU dataset is annotated to evaluate TSA, and extensive experiments show that our proposed TSA can significantly outperform several state-of-the-art baselines, justifying the effectiveness of TSA to ChatASU and the importance of considering the coreference and hallucination issues in ChatASU.

Weakly-supervised Phrase Grounding (WPG) is an emerging task of inferring the fine-grained phrase-region matching, while merely leveraging the coarse-grained sentence-image pairs for training. However, existing studies on WPG largely ignore the implicit phrase-region matching relations, which are crucial for evaluating the capability of models in understanding the deep multimodal semantics. To this end, this paper proposes an Implicit-Enhanced Causal Inference (IECI) approach to address the challenges of modeling the implicit relations and highlighting them beyond the explicit. Specifically, this approach leverages both the intervention and counterfactual techniques to tackle the above two challenges respectively. Furthermore, a high-quality implicit-enhanced dataset is annotated to evaluate IECI and detailed evaluations show the great advantages of IECI over the state-of-the-art baselines. Particularly, we observe an interesting finding that IECI outperforms the advanced multimodal LLMs by a large margin on this implicit-enhanced dataset, which may facilitate more research to evaluate the multimodal LLMs in this direction.

Previous work has demonstrated that, in the Variance Preserving (VP) scenario, the nascent Directly Denoising Diffusion Models (DDDM) can generate high-quality images in one step while achieving even better performance in multistep sampling. However, the Pseudo-LPIPS loss used in DDDM leads to concerns about the bias in assessment. Here, we propose a unified DDDM (uDDDM) framework that generates images in one-step/multiple steps for both Variance Preserving (VP) and Variance Exploding (VE) cases. We provide theoretical proofs of the existence and uniqueness of the model's solution paths, as well as the non-intersecting property of the sampling paths. Additionally, we propose an adaptive Pseudo-Huber loss function to balance the convergence to the true solution and the stability of convergence process.Through a comprehensive evaluation, we demonstrate that uDDDMs achieve FID scores comparable to the best-performing methods available for CIFAR-10 in both VP and VE. Specifically, uDDDM achieves one-step generation on CIFAR10 with FID of 2.63 and 2.53 for VE and VP respectively. By extending the sampling to 1000 steps, we further reduce FID score to 1.71 and 1.65 for VE and VP respectively, setting state-of-the-art performance in both cases.

In this paper, we present the Directly Denoising Diffusion Model (DDDM): a simple and generic approach for generating realistic images with few-step sampling, while multistep sampling is still preserved for better performance. DDDMs require no delicately designed samplers nor distillation on pre-trained distillation models. DDDMs train the diffusion model conditioned on an estimated target that was generated from previous training iterations of its own. To generate images, samples generated from the previous time step are also taken into consideration, guiding the generation process iteratively. We further propose Pseudo-LPIPS, a novel metric loss that is more robust to various values of hyperparameter. Despite its simplicity, the proposed approach can achieve strong performance in benchmark datasets. Our model achieves FID scores of 2.57 and 2.33 on CIFAR-10 in one-step and two-step sampling respectively, surpassing those obtained from GANs and distillation-based models. By extending the sampling to 1000 steps, we further reduce FID score to 1.79, aligning with state-of-the-art methods in the literature. For ImageNet 64x64, our approach stands as a competitive contender against leading models.

In this report, we present the 1st place solution for ICCV 2023 OmniObject3D Challenge: Sparse-View Reconstruction. The challenge aims to evaluate approaches for novel view synthesis and surface reconstruction using only a few posed images of each object. We utilize Pixel-NeRF as the basic model, and apply depth supervision as well as coarse-to-fine positional encoding. The experiments demonstrate the effectiveness of our approach in improving sparse-view reconstruction quality. We ranked first in the final test with a PSNR of 25.44614.

Surrounding perceptions are quintessential for safe driving for connected and autonomous vehicles (CAVs), where the Bird's Eye View has been employed to accurately capture spatial relationships among vehicles. However, severe inherent limitations of BEV, like blind spots, have been identified. Collaborative perception has emerged as an effective solution to overcoming these limitations through data fusion from multiple views of surrounding vehicles. While most existing collaborative perception strategies adopt a fully connected graph predicated on fairness in transmissions, they often neglect the varying importance of individual vehicles due to channel variations and perception redundancy. To address these challenges, we propose a novel Priority-Aware Collaborative Perception (PACP) framework to employ a BEV-match mechanism to determine the priority levels based on the correlation between nearby CAVs and the ego vehicle for perception. By leveraging submodular optimization, we find near-optimal transmission rates, link connectivity, and compression metrics. Moreover, we deploy a deep learning-based adaptive autoencoder to modulate the image reconstruction quality under dynamic channel conditions. Finally, we conduct extensive studies and demonstrate that our scheme significantly outperforms the state-of-the-art schemes by 8.27% and 13.60%, respectively, in terms of utility and precision of the Intersection over Union.

Ling Qin,1 Dongli Gao,1 Qian Wang,1 Xuewei Zheng,2 Jingjing Wang,2 Xingang Chen,2 Dongliao Fu,2 Haodi Ma,2 Junjia Tan,3 Qinan Yin2 1The First Affiliated Hospital, and College of Clinical Medicine of Henan University of Science and Technology, Luoyang, People’s Republic of China; 2Laboratory for Precision Medicine, School of Medical Technology and Engineering, Henan University of Science and Technology, Luoyang, People’s Republic of China; 3Section for Statistical Analysis and Data Collection, Luoyang Branch, Bank of China, Luoyang, People’s Republic of ChinaCorrespondence: Qinan Yin, School of Medical Technology and Engineering, Henan University of Science and Technology, No. 263 Kaiyuan Avenue, Luoyang, 471000, People’s Republic of China, Tel +86-13939928711, Email qinanyin@haust.edu.cnAbstract: ABO blood group antigens exhibit alternative phenotypes and genetically derived structures that are located on the red cell surface. The role of ABO blood group in cancer biology has been intensely reported by several studies, and it is now widely recognized that ABO antigens are associated with the risk and prognosis of several types of tumors, namely gastric cancer and pancreatic cancer. However, there have been contentious limited issues with the association between the ABO blood group and lymphoma. In this narrative review, based on literature data, we discuss the role of ABO blood group in the risk and prognosis of lymphoma and summarize the current knowledge of the underlying pathogenic mechanisms of the association. The possible association of ABO blood group with racial disparities and pathological classification in lymphoma patients is also discussed.Keywords: ABO blood group, lymphoma, survival, disease susceptibility, race

Denitrifying woodchip bioreactors (DWBs) have proven to be an efficient nature-based solution for nitrate removal. Modeling DWBs is required for improving their design and operation, but is hindered by the complexity of the modeled system where numerous chemical species and model parameters are needed. Reactions inside the woodchips are different from those at the edges, causing chemical localization (i.e., apparent simultaneous occurrence of incompatible reactions). We used the Multi Rate Mass Transfer (MRMT) approach to overcome these problems when simulating reactive transport processes in a DWB located at Kiruna, Sweden. Besides denitrification, other nitrogen-cycling processes (e.g., nitrification, dissimilatory nitrate reduction to ammonium, anammox) and alternative electron donors (e.g. oxygen, sulfate) were also considered. Biomass concentration is incorporated into the biochemical reaction rates, including growth and decay, to characterize microbial catalysis. We found that the MRMT model: 1) can account for the heterogeneity of the porous woodchips; 2) was capable of reproducing the nitrogen species evolution in the DWB with kinetic parameters from the literature; and 3) allows reproducing localized biochemical reactions (e.g., aerobic reactions on the woodchip edges, near the DWB entrance and anaerobic reactions inside); and 4) reproduces the full denitrification reactions sequence, but with the different reactions occurring in different portions of the woodchip (e.g., nitrate to nitrite near the edges and nitrite to nitrous oxide further inside). The latter observation suggests that increasing woodchip size may reduce the outflow of these undesired species., NITREM

Denitrifying woodchip bioreactors (DWBs) have proven to be an efficient nature-based solution for nitrate removal. Modeling DWBs is required for improving their design and operation, but is hindered by the complexity of the modeled system where numerous chemical species and model parameters are needed. Reactions inside the woodchips are different from those at the edges, causing chemical localization (i.e., apparent simultaneous occurrence of incompatible reactions). We used the Multi Rate Mass Transfer (MRMT) approach to overcome these problems when simulating reactive transport processes in a DWB located at Kiruna, Sweden. Besides denitrification, other nitrogen-cycling processes (e.g., nitrification, dissimilatory nitrate reduction to ammonium, anammox) and alternative electron donors (e.g. oxygen, sulfate) were also considered. Biomass concentration is incorporated into the biochemical reaction rates, including growth and decay, to characterize microbial catalysis. We found that the MRMT model: 1) can account for the heterogeneity of the porous woodchips; 2) was capable of reproducing the nitrogen species evolution in the DWB with kinetic parameters from the literature; and 3) allows reproducing localized biochemical reactions (e.g., aerobic reactions on the woodchip edges, near the DWB entrance and anaerobic reactions inside); and 4) reproduces the full denitrification reactions sequence, but with the different reactions occurring in different portions of the woodchip (e.g., nitrate to nitrite near the edges and nitrite to nitrous oxide further inside). The latter observation suggests that increasing woodchip size may reduce the outflow of these undesired species., The authors acknowledge the financial support of projects NITREM, RESTORA (Agència Catalana de l'Aigua, ACA210/18/0040), and ConMimo (Spanish Research Agency, AEI, number TED2021-131188B-C31). NITREM (project number 17013) has received funding from the European Institute of Innovation and Technology (EIT). This body of the European Union received support from the European Union's Horizon 2020 research and innovation programme. Additional funding was obtained from the Generalitat de Catalunya (2017 SGR1485) and the Spanish Ministry of Science and Innovation (Centre of Excellence Severo-Ochoa, CSIC-IDAEA, CEX2018-000794-S)., Peer Reviewed, Postprint (published version)

Patient-derived tumor xenograft (PDX)/organoid (PDO), driven by cancer stem cells (CSC), are considered the most predictive models for translational oncology. Large PDX collections reflective of patient populations have been created and used extensively to test various investigational therapies, including population-trials as surrogate subjects in vivo. PDOs are recognized as in vitro surrogates for patients amenable for high-throughput screening (HTS). We have built a biobank of carcinoma PDX-derived organoids (PDXOs) by converting an existing PDX library and confirmed high degree of similarities between PDXOs and parental PDXs in genomics, histopathology and pharmacology, suggesting "biological equivalence or interchangeability" between the two. Here we demonstrate the applications of PDXO biobank for HTS "matrix" screening for both lead compounds and indications, immune cell co-cultures for immune-therapies and engineering enables in vitro/in vivo imaging. This large biobank of >550 matched pairs of PDXs/PDXOs across different cancers could become powerful tools for the future cancer drug discovery.

Yang G, Zhang P, Li L, et al. Diabetes Metab Syndr Obes. 2023;16:1515–1523 The authors have advised there is an error in the affiliation on page 1515. The affiliation should read “Zhengzhou Central Hospital Affiliated to Zhengzhou University, Zhengzhou, Henan, People’s Republic of China”.

Jingjing Wang,1 Lin Wang,1 Zhe Pang,1 Qingmiao Ge,1 Yonggui Wu,1,2 Xiangming Qi1 1Department of Nephropathy, The First Affiliated Hospital of Anhui Medical University, Hefei, Anhui, 230022, People’s Republic of China; 2Center for Scientific Research of Anhui Medical University, Hefei, Anhui, 230022, People’s Republic of ChinaCorrespondence: Yonggui Wu; Xiangming Qi, Department of Nephropathy, The First Affiliated Hospital of Anhui Medical University, Hefei, Anhui, People’s Republic of China, Tel +86 551 6292 2111, Email wuyonggui@medmail.com.cn; qxm119@126.comPurpose: Diabetic kidney disease (DKD) is the leading cause of chronic kidney disease (CKD) worldwide. Elucidation of the molecular mechanisms underlying ferroptosis and immunity in DKD could aid the development of potentially effective therapeutics. This study aimed to perform an integrated analysis of ferroptosis and immune-related differentially expressed mRNAs (DEGs) in DKD.Materials and Methods: Gene expression profiles of samples obtained from patients with DKD and controls were downloaded from the Gene Expression Omnibus (GEO) database. The potential differentially expressed genes (DEGs) were screened using R software, and ferroptosis immune-related differentially expressed genes (FIRDEGs) were extracted from the DEGs. We performed functional enrichment analyses, and constructed protein–protein interaction (PPI) networks, transcription factor (TFs)-gene networks, and gene-drug networks to explore their potential biological functions. Correlation analysis and receiver operating characteristic curves were used for evaluating the FIRDEGs. We used the CIBERSORT algorithm to examine the composition of immune cells and determine the relationship between FIRDEG signatures and immune cells. Finally, the RNA expression of six FIRDEGs was validated in animal kidney samples using RT-PCR.Results: We identified 80 FIRDEGs and performed their functional analyses. We identified six hub genes (Ccl5, Il18, Cybb

Haoyuan Xu,&ast; Wanpeng Li,&ast; Xiaole Song,&ast; Huankang Zhang, Huan Wang, Jingjing Wang, Li Hu, Houyong Li, Xicai Sun,&ast; Dehui Wang&ast; Department of Otolaryngology-Head and Neck Surgery, Affiliated Eye Ear Nose and Throat Hospital, Fudan University, Shanghai, People’s Republic of China&ast;These authors contributed equally to this workCorrespondence: Dehui Wang; Xicai Sun, Department of Otolaryngology-Head and Neck Surgery, Affiliated Eye Ear Nose and Throat Hospital, Fudan University, 83 Fen Yang Road, Shanghai, People’s Republic of China, Email wangdehuient@sina.com; laryngeal@163.comBackground: As a cancer stem cells (CSCs) surface marker, Lgr5 plays an important role in the signal transduction of cancer cells and is a potential biomarker for cancer diagnosis and prognosis. However, the expression and prognostic value of Lgr5 in recurrent nasopharyngeal carcinoma (rNPC) remains ambiguous.Materials: We used RNA sequencing to screen differentially expressed mRNAs in eleven specimens of rNPC tissues and five fresh adjacent normal tissue samples and the CSC marker, Lgr5, was identified. The expression level of Lgr5 in rNPC samples was also detected by immunohistochemistry and Western blot assay. The chi-square test was used to analyze the relationship between the clinicopathological variables and the immunostaining of Lgr5. The Log-rank method was used for prognosis analysis. The Cox regression model was used for univariate and multivariate analysis.Results: Significantly elevated expression of Lgr5 in the rNPC tissues was observed compared to the normal tissues using RNA sequencing, Western blot and immunohistochemistry. The expression of Lgr5 was significantly correlated with the T stage (P=0.014). High Lgr5 expression (P=0.007), tumor necrosis (P=0.013) and WHO type II (P=0.043) in rNPC patients exhibited worse overall survival (OS). Lgr5 expression was proved to be an independent risk factor for OS (P=0.035) in multivariate a

Hole transport materials (HTMs) have greatly advanced the progress of solution-based electronic devices in the past few years. Nevertheless, most devices employing dopant-free organic HTMs can only deliver inferior performance. In this work, we introduced a novel "polymer synergy" strategy to develop versatile dopant-free polymer HTMs for quantum dot/perovskite solar cells and photodetectors. With this synergy strategy, the optical, electrical and aggregation properties of polymer HTMs can be modulated, resulting in complementary absorption, high hole mobility, favorable energy landscape and moderate aggregation. Moreover, a clear orientational transition was observed for the developed HTMs with a 9-fold increase in the face-on/edge-on ratio, providing a highway-like carrier transport for electronic devices, as revealed by in situ characterization and ultrafast transient absorption. With these benefits, the photovoltaic and photodetection performance of quantum dot devices were boosted from 11.8% to 13.5% and from 2.95 x 10(12) to 3.41 x 10(13) Jones (over a 10-fold increase), respectively. Furthermore, the developed polymer HTMs can also significantly enhance the photovoltaic and photodetection performance of perovskite devices from 15.1% to 22.7% and from 2.7 x 10(12) to 2.17 x 10(13)Jones with the same device structure, indicating their great application potential in the emerging optoelectronics.

Guanyu Yang,&ast; Pin Zhang,&ast; Liumei Li, Jingjing Wang, Pengfei Jiao, Jie Wang, Qinjun Chu Department of Anesthesiology and Perioperative Medicine, Zhengzhou Central Hospital, Zhengzhou, Henan, People’s Republic of China&ast;These authors contributed equally to this workCorrespondence: Qinjun Chu, Tongbai North Road 16, Zhongyuan District, Zhengzhou, Henan, People’s Republic of China, Tel/Fax +86 13643711142, Email mzk123456@zzu.edu.cnPurpose: This study aims to compare the conventional lung protective ventilation strategy (LPVS) with driving pressure-guided ventilation in obese patients undergoing laparoscopic sleeve gastrectomy (LSG).Methods: Forty-five patients undergoing elective LSG under general anesthesia were randomly assigned to the conventional LPVS group (group L) or the driving pressure-guided ventilation group (group D) using random numbers generated by Excel. The primary outcome was the driving pressure of both groups 90 min after pneumoperitoneum.Results: After 30 min of pneumoperitoneum, 90 min of pneumoperitoneum, 10 min of closing the pneumoperitoneum, and restoring the supine position, the driving pressure of group L and group D were 20.0 ± 2.9 cm H2O vs 16.6 ± 3.0 cm H2O (P < 0.001), 20.7 ± 3.2 cm H2O vs 17.3 ± 2.8 cm H2O (P < 0.001), and 16.3 ± 3.1 cm H2O vs 13.3 ± 2.5 cm H2O (P = 0.001), respectively; the respiratory compliance of groups L and D were 23.4 ± 3.7 mL/cm H2O vs 27.6 ± 5.1 mL/cm H2O (P = 0.003), 22.7 ± 3.8 mL/cm H2O vs 26.4 ± 3.5 mL/cm H2O (P = 0.005), and 29.6 ± 6.8 mL/cm H2O vs 34.7 ± 5.3 mL/cm H2O (P = 0.007), respectively. The intraoperative PEEP in groups L and group D was 5 (5– 5) cm H2O vs 10 (9– 11) cm H2O (P < 0.001).Conclusion: An individualized peep-based driving pressure-guided ventilation strategy can reduce intraoperative driving pressure and increase respiratory compliance in obese patients undergoing LSG.Keywords: lung protective ventilation strategy, driving pressure-guid

The proliferation of users and data traffic poses substantial pressure on resource management in the Internet of Things (IoT). In addition to beneficially allocating scarce network resources, it also needs to meet differentiated users' Quality-of-Service (QoS) requirements, such as low delay, high security, etc. The distributed management architecture of blockchain and its inherent security features bring inspiration to resource management in the IoT. In this article, we propose a blockchain-enabled resource orchestration scheme for IoT by deep reinforcement learning (DRL), where the IoT edge server and the end user can reach a consensus on the allocation of network resources based on blockchain theory. Moreover, relying on the policy network, the intelligent agent can be trained by these resource attributes to fully perceive the change of the network's state and hence make dynamic resource allocation decisions. Finally, simulation results show that the proposed resource orchestration scheme has good performance in comparison to other security resource allocation algorithms. The average revenue, the user request acceptance rate, and the profitability are increased by an average of 8.5%, 1.8%, and 11.9%, respectively, compared with other algorithms.

The compelling applications of Low earth orbit (LEO) satellite networks in our daily lives have been witnessed in recent years, ranging from weather forecasts to military monitoring. LEO satellite networks have grown in importance as a complement to terrestrial networks aiming at delivering global, ubiquitous communication. However, due to the LEO network topology keeps changing dynamically, the development of efficient routing algorithms becomes one of the challenges in the LEO network. Traditional routing policies hardly solve the rerouting problem caused by link switchover and the high calculation cost caused by large-scale irregular satellite topology. In this paper, we propose a knowledge graph-aided representation of satellite network topologies and routing architecture to optimize path selection and calculation cost for lower packet loss ratio and average delay. Moreover, we generate a routing policy by predicting potential relations between data packets and nodes to select the best relay nodes with the maximum probability of forwarding relations. Finally, extensive simulations are performed to evaluate the performance and availability of our proposed algorithm. © 1967-2012 IEEE.

Although metal and metalloid concentrations in wildfire ashes have been documented, the nature and concen-trations of incidental nanomaterials (INMs) in wildland-urban interface (WUI) fire ashes have received consid-erably less attention. In this study, the total metal and metalloid concentrations of 57 vegetation, structural, and vehicle ashes and underlying soils collected at the WUI following the 2020 fire season in northern California - North Complex Fire and LNU Lightning Complex Fire - were determined using inductively coupled plasma-time of flight-mass spectrometry after microwave-assisted acid digestion. The concentrations of Ti, Zn, Cu, Ni, Pb, Sn, Sb, Co, Bi, Cr, Ba, As, Rb, and W are generally higher in structural/vehicle-derived ashes than in vegetation -derived ashes and soils. The concentrations of Ca, Sr, Rb, and Ag increased with increased combustion completeness (e.g., black ash < gray ash < white ash), whereas those of C, N, Zn, Pb, and In decreased with increased combustion completeness. The concentration of anthropogenic Ti - determined by mass balance calculations and shifts in Ti/Nb above the natural background ratios - was highest in vehicle ash (median: 30.8 g kg -1, range: 4.5-41.0 g kg -1) followed by structural ash (median: 5.5 g kg -1, range: of 0-77.4 g kg -1). Various types of carbonaceous INM (e.g., amorphous carbon, turbostratic-like carbon, and carbon associated with zinc oxides) and metal-bearing INMs (e.g., Ti, Cu, Fe, Zn, Mn, Pb, and Cr) with sizes between few nanometers to few hundreds of nanometers were evidenced in ashes using transmission electron microscopy, including energy dispersive X-ray spectroscopy. Overall, this study demonstrates the abundance of a variety of metals and met-alloids in the form of INMs in WUI fire ashes. This study also highlights the need for further research into the formation, transformation, reactivity, fate, and effects of INMs during and following fires at the WUI.

In addition to the combustion of vegetation, fires at the wildland-urban interface (WUI) burn structural mate-rials, including chromated copper arsenate (CCA)-treated wood. This study identifies, quantifies, and charac-terizes Cr-, Cu-, and As-bearing incidental nanomaterials (INMs) in WUI fire ashes collected from three residential structures suspected to have originated from the combustion of CCA-treated wood. The total elemental concentrations were determined by inductively coupled plasma-time of flight-mass spectrometry (ICP-TOF-MS) following acid digestion. The crystalline phases were determined using transmission electron micro-scopy (TEM), specifically using electron diffraction and high-resolution imaging. The multi-element single particle composition and size distribution were determined by single particle (SP)-ICP-TOF-MS coupled with agglomerative hierarchical clustering analysis. Chromium, Cu, and As are the dominant elements in the ashes and together account for 93%, 83%, and 24% of the total mass of measured elements in the ash samples. Chromium, Cu, and As phases, analyzed by TEM, most closely match CrO3, CrO2, eskolaite (Cr2O3), CuCrO2, CuCr2O4, CrAs2O6, As2O5, AsO2, claudetite (As2O3, monoclinic), or arsenolite (As2O3, cubic), although a bona fide phase identification for each particle was not always possible. These phases occur predominantly as het-eroaggregates. Multi-element single particle analyses demonstrate that Cr occurs as a pure phase (i.e., Cr oxides) as well as in association with other elements (e.g., Cu and As); Cu occurs predominantly in association with Cr and As; and As occurs as As oxides and in association with Cu and Cr. Several Cr, Cu, and As clusters were identified and the molar ratios of Cr/Cu and Cr/As within these clusters are consistent with the crystalline phases identified by TEM as well as their heteroaggregates. These results indicate that WUI fires can lead to significant release of CCA constituents and their combustio

Denitrifying woodchip bioreactors (DWBs) have proven to be an efficient nature-based solution for nitrate removal. Modeling DWBs is required for improving their design and operation, but is hindered by the complexity of the modeled system where numerous chemical species and model parameters are needed. Reactions inside the woodchips are different from those at the edges, causing chemical localization (i.e., apparent simultaneous occurrence of incompatible reactions). We used the Multi Rate Mass Transfer (MRMT) approach to overcome these problems when simulating reactive transport processes in a DWB located at Kiruna, Sweden. Besides denitrification, other nitrogen-cycling processes (e.g., nitrification, dissimilatory nitrate reduction to ammonium, anammox) and alternative electron donors (e.g. oxygen, sulfate) were also considered. Biomass concentration is incorporated into the biochemical reaction rates, including growth and decay, to characterize microbial catalysis. We found that the MRMT model: 1) can account for the heterogeneity of the porous woodchips; 2) was capable of reproducing the nitrogen species evolution in the DWB with kinetic parameters from the literature; and 3) allows reproducing localized biochemical reactions (e.g., aerobic reactions on the woodchip edges, near the DWB entrance and anaerobic reactions inside); and 4) reproduces the full denitrification reactions sequence, but with the different reactions occurring in different portions of the woodchip (e.g., nitrate to nitrite near the edges and nitrite to nitrous oxide further inside). The latter observation suggests that increasing woodchip size may reduce the outflow of these undesired species.

Iris presentation attack detection (PAD) has achieved great success under intra-domain settings but easily degrades on unseen domains. Conventional domain generalization methods mitigate the gap by learning domain-invariant features. However, they ignore the discriminative information in the domain-specific features. Moreover, we usually face a more realistic scenario with only one single domain available for training. To tackle the above issues, we propose a Single Domain Dynamic Generalization (SDDG) framework, which simultaneously exploits domain-invariant and domain-specific features on a per-sample basis and learns to generalize to various unseen domains with numerous natural images. Specifically, a dynamic block is designed to adaptively adjust the network with a dynamic adaptor. And an information maximization loss is further combined to increase diversity. The whole network is integrated into the meta-learning paradigm. We generate amplitude perturbed images and cover diverse domains with natural images. Therefore, the network can learn to generalize to the perturbed domains in the meta-test phase. Extensive experiments show the proposed method is effective and outperforms the state-of-the-art on LivDet-Iris 2017 dataset., Comment: ICASSP 2023 Camera Ready

Multi-sensor clues have shown promise for object segmentation, but inherent noise in each sensor, as well as the calibration error in practice, may bias the segmentation accuracy. In this paper, we propose a novel approach by mining the Cross-Modal Semantics to guide the fusion and decoding of multimodal features, with the aim of controlling the modal contribution based on relative entropy. We explore semantics among the multimodal inputs in two aspects: the modality-shared consistency and the modality-specific variation. Specifically, we propose a novel network, termed XMSNet, consisting of (1) all-round attentive fusion (AF), (2) coarse-to-fine decoder (CFD), and (3) cross-layer self-supervision. On the one hand, the AF block explicitly dissociates the shared and specific representation and learns to weight the modal contribution by adjusting the \textit{proportion, region,} and \textit{pattern}, depending upon the quality. On the other hand, our CFD initially decodes the shared feature and then refines the output through specificity-aware querying. Further, we enforce semantic consistency across the decoding layers to enable interaction across network hierarchies, improving feature discriminability. Exhaustive comparison on eleven datasets with depth or thermal clues, and on two challenging tasks, namely salient and camouflage object segmentation, validate our effectiveness in terms of both performance and robustness. The source code is publicly available at https://github.com/Zongwei97/XMSNet., Comment: ACM MM 2023

As an evolving successor to the mobile Internet, the extended reality (XR) devices can generate a fully digital immersive environment similar to the real world, integrating integrating virtual and real-world elements. However, in addition to the difficulties encountered in traditional communications, there emerge a range of new challenges such as ultra-massive access, real-time synchronization as well as unprecedented amount of multi-modal data transmission and processing. To address these challenges, semantic communications might be harnessed in support of XR applications, whereas it lacks a practical and effective performance metric. For broadening a new path for evaluating semantic communications, in this paper, we construct a multi-user uplink non-orthogonal multiple access (NOMA) system to analyze its transmission performance by harnessing a novel metric called age of incorrect information (AoII). First, we derive the average semantic similarity of all the users based on DeepSC and obtain the closed-form expressions for the packets' age of information (AoI) relying on queue theory. Besides, we formulate a non-convex optimization problem for the proposed AoII which combines both error-and AoI-based performance under the constraints of semantic rate, transmit power and status update rate. Finally, in order to solve the problem, we apply an exact linear search based algorithm for finding the optimal policy. Simulation results show that the AoII metric can beneficially evaluate both the error- and AoI-based transmission performance simultaneously.

As an evolving successor to the mobile Internet, the Metaverse creates the impression of an immersive environment, integrating the virtual as well as the real world. In contrast to the traditional mobile Internet based on servers, the Metaverse is constructed by billions of cooperating users by harnessing their smart edge devices having limited communication and computation resources. In this immersive environment an unprecedented amount of multi-modal data has to be processed. To circumvent this impending bottleneck, low-rate semantic communication might be harnessed in support of the Metaverse. But given that private multi-modal data is exchanged in the Metaverse, we have to guard against security breaches and privacy invasions. Hence we conceive a trust-worthy semantic communication system for the Metaverse based on a federated learning architecture by exploiting its distributed decision-making and privacy-preserving capability. We conclude by identifying a suite of promising research directions and open issues.

Most existing approaches for point cloud normal estimation aim to locally fit a geometric surface and calculate the normal from the fitted surface. Recently, learning-based methods have adopted a routine of predicting point-wise weights to solve the weighted least-squares surface fitting problem. Despite achieving remarkable progress, these methods overlook the approximation error of the fitting problem, resulting in a less accurate fitted surface. In this paper, we first carry out in-depth analysis of the approximation error in the surface fitting problem. Then, in order to bridge the gap between estimated and precise surface normals, we present two basic design principles: 1) applies the $Z$-direction Transform to rotate local patches for a better surface fitting with a lower approximation error; 2) models the error of the normal estimation as a learnable term. We implement these two principles using deep neural networks, and integrate them with the state-of-the-art (SOTA) normal estimation methods in a plug-and-play manner. Extensive experiments verify our approaches bring benefits to point cloud normal estimation and push the frontier of state-of-the-art performance on both synthetic and real-world datasets., Comment: The first two authors contributed equally to this work. The source code are available at https://github.com/hikvision-research/3DVision. Accepted to CVPR 2023

Zinc acts as an important cofactor in the body and is essential for normal functions. Several zinc interventions have been implemented worldwide to improve the public's zinc status, but limited studies have assessed their cost-effectiveness. To help inform decision-making on zinc interventions to maximize benefits within a fixed budget, we took China as an example and evaluated the cost-effectiveness of three interventions, that is, supplementation, food fortification, and biofortification. As an essential group at high risk of zinc deficiency, children aged 5–14 years, who account for 10% of the Chinese population, were selected as the target group in this study. We constructed a decision-analytic Markov model to determine the cost-effectiveness of interventions in China under different scenarios. In our model, biofortification through conventional breeding was shown to be the most cost-effective approach in most scenarios. Compared with other interventions, zinc supplementation gained fewer quality-adjusted life years at a higher net cost, suggesting that this common approach may not be optimal for large-scale, long-term implementation at the national level. While the robustness of the results was further confirmed by the sensitivity analysis, more research is needed to assess the cost-effectiveness of addressing zinc deficiency with other interventions. Further clinical trials are also expected to evaluate the effectiveness of zinc interventions in reducing pneumonia cases [EconLit Citations: I18, Q16, Q18].

Large Language Models (LLMs), representing a significant achievement in artificial intelligence (AI) research, have demonstrated their ability in a multitude of tasks. This project aims to explore the capabilities of GPT-3.5, a leading example of LLMs, in processing the sentiment analysis of Internet memes. Memes, which include both verbal and visual aspects, act as a powerful yet complex tool for expressing ideas and sentiments, demanding an understanding of societal norms and cultural contexts. Notably, the detection and moderation of hateful memes pose a significant challenge due to their implicit offensive nature. This project investigates GPT's proficiency in such subjective tasks, revealing its strengths and potential limitations. The tasks include the classification of meme sentiment, determination of humor type, and detection of implicit hate in memes. The performance evaluation, using datasets from SemEval-2020 Task 8 and Facebook hateful memes, offers a comparative understanding of GPT responses against human annotations. Despite GPT's remarkable progress, our findings underscore the challenges faced by these models in handling subjective tasks, which are rooted in their inherent limitations including contextual understanding, interpretation of implicit meanings, and data biases. This research contributes to the broader discourse on the applicability of AI in handling complex, context-dependent tasks, and offers valuable insights for future advancements.

The explosive growth of smart devices (e.g., mobile phones, vehicles, drones) with sensing, communication, and computation capabilities gives rise to an unprecedented amount of data. The generated massive data together with the rapid advancement of machine learning (ML) techniques spark a variety of intelligent applications. To distill intelligence for supporting these applications, federated learning (FL) emerges as an effective distributed ML framework, given its potential to enable privacy-preserving model training at the network edge. In this article, we discuss the challenges and solutions of achieving scalable wireless FL from the perspectives of both network design and resource orchestration. For network design, we discuss how task-oriented model aggregation affects the performance of wireless FL, followed by proposing effective wireless techniques to enhance the communication scalability via reducing the model aggregation distortion and improving the device participation. For resource orchestration, we identify the limitations of the existing optimization-based algorithms and propose three task-oriented learning algorithms to enhance the algorithmic scalability via achieving computation-efficient resource allocation for wireless FL. We highlight several potential research issues that deserve further study., Comment: This paper has been accepted by IEEE Internet of Things Magazine

The Internet of underwater things (IoUT) is envisioned to be an essential part of maritime activities. Given the IoUT devices' wide-area distribution and constrained transmit power, autonomous underwater vehicles (AUVs) have been widely adopted for collecting and forwarding the data sensed by IoUT devices to the surface-stations. In order to accommodate the diverse requirements of IoUT applications, it is imperative to conceive a multi-tier underwater computing (MTUC) framework by carefully harnessing both the computing and the communications as well as the storage resources of both the surface-station and of the AUVs as well as of the IoUT devices. Furthermore, to meet the stringent energy constraints of the IoUT devices and to reduce the operating cost of the MTUC framework, a joint environment-aware AUV trajectory design and resource management problem is formulated, which is a high-dimensional NP-hard problem. To tackle this challenge, we first transform the problem into a Markov decision process (MDP) and solve it with the aid of the asynchronous advantage actor-critic (A3C) algorithm. Our simulation results demonstrate the superiority of our scheme., Comment: Accepted by IEEE Journal on Selected Areas in Communications

Aim of study: To explore the different patterns of fine-root traits by elucidating changes in their chemistries and morphologies in Chinese fir plantations along an altitudinal gradient. Area of study: National Mazongling Nature Reserve (Anhui Province). Material and methods: Soil and fine roots (≤ 2 mm) samples were extracted from three soil layers (0-10 cm, 10-20 cm, and 20-30 cm) at four altitudes (750 m, 850 m, 1000 m, and 1150 m), after which their nutrient concentrations and morphological traits, respectively, were quantified. We employed mixed model ANOVA to test the effects of altitude, soil layer, and their interactions on the characteristics of soil and fine roots. The relationships between the functional traits of fine roots and climate, soil and stand structures were evaluated by the standard major axis regression and the structural equation model. Main results: The chemical traits of fine roots were higher at medium altitudes (which decreased significantly with the soil layer). In contrast the morphological traits of fine roots did not change significantly. In chemical traits, both altitude, organic matter components, and soil total phosphorus (TP) exerted dominant effects on fine-root N, and both altitude and soil TP exerted dominant effects on fine-root P. However, in morphological traits, we found that altitude and soil C:N were crucial impact factors. Research highlights: Fine roots might preferentially adjust their chemical traits rather than morphological traits to facilitate higher root efficiencies in response to variable environmental conditions.

Jun Ma,1,* Hongcheng Liu,2,* Jingjing Wang,3,* Wenting Li,1 Lin Fan,1 Wenwen Sun1 1Department of Tuberculosis and Shanghai Key Laboratory of Tuberculosis, Shanghai Pulmonary Hospital, School of Medicine, Tongji University, Shanghai, People’s Republic of China; 2Department of Thoracic Surgery, Shanghai Pulmonary Hospital, School of Medicine, Tongji University, Shanghai, People’s Republic of China; 3Department of Emergency, Shanghai Pulmonary Hospital, School of Medicine, Tongji University, Shanghai, People’s Republic of China*These authors contributed equally to this workCorrespondence: Wenwen Sun; Lin Fan, Department of Tuberculosis and Shanghai Key Laboratory of Tuberculosis, Shanghai Pulmonary Hospital, School of Medicine, Tongji University, Shanghai, 200433, People’s Republic of China, Email sunwenwen_1583@163.com; fanlinsj@163.comBackground: China is the region with a high global burden of rifampicin resistance/multidrug-resistant tuberculosis (RR/MDR-TB) and low HIV incidence. Our aim was to assess the clinical and demographic characteristics of RR/MDR-extrapulmonary tuberculosis (EPTB) from 2015 to 2019 to provide evidence for the prevention and control of the disease in high TB burden areas.Methods: We investigated the clinical and demographic data of all MDR/RR-EPTB cases in a TB specialized hospital from China and compared the cases with the MDR/RR-pulmonary tuberculosis (PTB) patients over the same period.Results: Of the RR/MDR-TB patients enrolled, 15.4 were EPTB. The most common anatomical site was pleural/chest wall (25.9%). Compared with RR/MDR-PTB, females were more likely to be susceptible to RR/MDR-EPTB (OR 1.65, 95% CI 1.52– 1.77); the risk of RR/MDR-EPTB for 25– 44 years group increased (OR 1.61, 95% CI 1.52– 1.77), and then decreased with the increasing age (OR 1.48, 95% CI 0.74– 1.69 for 44– 65 years group and OR 2.23, 95% CI 0.98– 2.71 for &g

The presence of low permeability regions within porous media impacts solute transport and the distribution of species concentrations. Therefore, (bio)chemical reactions are equally affected. Multirate Mass Transfer (MRMT) models can be used to represent this anomalous transport process. MRMT conceptualizes the medium as a set of multiple continua: one mobile zone and multiple immobile zones. It simulates species transport in mobile and immobile zones simultaneously, which are related by first-order mass exchange. Numerical modeling of reactive transport in this kind of multicontinua media is complex and demanding because of the high dimensionality of the problem. In this paper, we establish the governing equations of reactive transport in multicontinuum media incorporating chemical kinetics into the governing equations. We propose a general numerical solution of reactive transport with MRMT by applying direct substitution approach (DSA) based on Newton-Raphson method. The efficiency of the proposed algorithm benefits of the block structure of the system, which allows us to eliminate immobile zones equations and leads to significant savings in CPU time. We test the validity of the developed solution by comparison with other numerical and analytical solutions., The authors acknowledge the support of project NITREM initiated and funded by EIT (European Institute of Innovation and Technology), and the Water-JPI European Union project MARadentro, project number PCI2019-103425-WW2017., Peer Reviewed, Postprint (published version)

The 6G will undergo an unprecedented transformation to revolutionize the wireless system evolution from connected things to connected intelligence. Additionally, data scattered around the industrial environments can be collected for the sake of enabling intelligent operations. In this paper, the promising multi-access edge computing (MEC) service is introduced into the IIoT system to assist the computation offloading and resource allocation for different compelling applications. Moreover, relying on defining a total cost function as a weighted sum of task delay and energy consumption, a novel deep reinforcement learning (DRL) based framework is proposed to jointly optimize task offloading and resource allocation. More specifically, the task offloading is decomposed with the aid of the new isotone action generation technique (IAGT) and adaptive action aggregation update strategy (3AUS) based on the proposed DRL framework, and the initial problem can be transformed into a convex optimization problem to solve the resource allocation for each IIoT device. Additionally, we periodically renovate the offloading policy in the DRL framework so that our proposed DRL-based decision-making algorithm can beneficially adapt to different network environments. Finally, extensive simulation demonstrate that our proposed algorithm for each IIoT device can obtain quasi-optimal system performance compared with some conventional baseline algorithms IEEE

In this work, we apply liquid cascade centrifugation to highly concentrated graphene dispersions produced by liquid-phase exfoliation in water with an insoluble bis-pyrene stabilizer to obtain fractions containing nanosheets with different lateral size distributions. The concentration, stability, size, thickness, and the cytotoxicity profile are studied as a function of the initial stabilizer concentration for each fraction. Our results show that there is a critical initial amount of stabilizer (0.4 mg/mL) above which the dispersions show reduced concentration, stability, and biocompatibility, no matter the lateral size of the flakes.

Gaze estimation is the fundamental basis for many visual tasks. Yet, the high cost of acquiring gaze datasets with 3D annotations hinders the optimization and application of gaze estimation models. In this work, we propose a novel Head-Eye redirection parametric model based on Neural Radiance Field, which allows dense gaze data generation with view consistency and accurate gaze direction. Moreover, our head-eye redirection parametric model can decouple the face and eyes for separate neural rendering, so it can achieve the purpose of separately controlling the attributes of the face, identity, illumination, and eye gaze direction. Thus diverse 3D-aware gaze datasets could be obtained by manipulating the latent code belonging to different face attributions in an unsupervised manner. Extensive experiments on several benchmarks demonstrate the effectiveness of our method in domain generalization and domain adaptation for gaze estimation tasks., Comment: 10 pages, 8 figures, submitted to CVPR 2023

Depth cues are known to be useful for visual perception. However, direct measurement of depth is often impracticable. Fortunately, though, modern learning-based methods offer promising depth maps by inference in the wild. In this work, we adapt such depth inference models for object segmentation using the objects' "pop-out" prior in 3D. The "pop-out" is a simple composition prior that assumes objects reside on the background surface. Such compositional prior allows us to reason about objects in the 3D space. More specifically, we adapt the inferred depth maps such that objects can be localized using only 3D information. Such separation, however, requires knowledge about contact surface which we learn using the weak supervision of the segmentation mask. Our intermediate representation of contact surface, and thereby reasoning about objects purely in 3D, allows us to better transfer the depth knowledge into semantics. The proposed adaptation method uses only the depth model without needing the source data used for training, making the learning process efficient and practical. Our experiments on eight datasets of two challenging tasks, namely camouflaged object detection and salient object detection, consistently demonstrate the benefit of our method in terms of both performance and generalizability., Comment: Accepted to ICCV 2023

A great variety of complex systems, from user interactions in communication networks to transactions in financial markets, can be modeled as temporal graphs consisting of a set of vertices and a series of timestamped and directed edges. Temporal motifs are generalized from subgraph patterns in static graphs which consider edge orderings and durations in addition to topologies. 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. Moreover, they cannot work in the streaming model where edges are observed incrementally over time. In this paper, we focus on approximate temporal motif counting via random sampling. We first propose two sampling algorithms for temporal motif counting in the offline setting. The first is an edge sampling (ES) algorithm for estimating the number of instances of any temporal motif. The second is an improved edge-wedge sampling (EWS) algorithm that hybridizes edge sampling with wedge sampling for counting temporal motifs with $3$ vertices and $3$ edges. Furthermore, we propose two algorithms to count temporal motifs incrementally in temporal graph streams by extending the ES and EWS algorithms referred to as SES and SEWS. We provide comprehensive analyses of the theoretical bounds and complexities of our proposed algorithms. Finally, we perform extensive experimental evaluations of our proposed algorithms on several real-world temporal graphs. The results show that ES and EWS have higher efficiency, better accuracy, and greater scalability than state-of-the-art sampling methods for temporal motif counting in the offline setting. Moreover, SES and SEWS achieve up to three orders of magnitude speedups over ES and EWS while having comparable estimation errors for temporal motif counting in the streaming setting., Comment: 27 pages, 11 figures; overlapped with arXiv:2007.14028

The rapid development of point cloud learning has driven point cloud completion into a new era. However, the information flows of most existing completion methods are solely feedforward, and high-level information is rarely reused to improve low-level feature learning. To this end, we propose a novel Feedback Network (FBNet) for point cloud completion, in which present features are efficiently refined by rerouting subsequent fine-grained ones. Firstly, partial inputs are fed to a Hierarchical Graph-based Network (HGNet) to generate coarse shapes. Then, we cascade several Feedback-Aware Completion (FBAC) Blocks and unfold them across time recurrently. Feedback connections between two adjacent time steps exploit fine-grained features to improve present shape generations. The main challenge of building feedback connections is the dimension mismatching between present and subsequent features. To address this, the elaborately designed point Cross Transformer exploits efficient information from feedback features via cross attention strategy and then refines present features with the enhanced feedback features. Quantitative and qualitative experiments on several datasets demonstrate the superiority of proposed FBNet compared to state-of-the-art methods on point completion task., Comment: The first two authors contributed equally to this work. The source code and model are available at https://github.com/hikvision-research/3DVision/. Accepted to ECCV 2022 as oral presentation

Point cloud upsampling focuses on generating a dense, uniform and proximity-to-surface point set. Most previous approaches accomplish these objectives by carefully designing a single-stage network, which makes it still challenging to generate a high-fidelity point distribution. Instead, upsampling point cloud in a coarse-to-fine manner is a decent solution. However, existing coarse-to-fine upsampling methods require extra training strategies, which are complicated and time-consuming during the training. In this paper, we propose a simple yet effective cascaded refinement network, consisting of three generation stages that have the same network architecture but achieve different objectives. Specifically, the first two upsampling stages generate the dense but coarse points progressively, while the last refinement stage further adjust the coarse points to a better position. To mitigate the learning conflicts between multiple stages and decrease the difficulty of regressing new points, we encourage each stage to predict the point offsets with respect to the input shape. In this manner, the proposed cascaded refinement network can be easily optimized without extra learning strategies. Moreover, we design a transformer-based feature extraction module to learn the informative global and local shape context. In inference phase, we can dynamically adjust the model efficiency and effectiveness, depending on the available computational resources. Extensive experiments on both synthetic and real-scanned datasets demonstrate that the proposed approach outperforms the existing state-of-the-art methods., Comment: The first two authors contributed equally to this work. The code is publicly available at https://github.com/hikvision-research/3DVision. Accepted to ACCV 2022 as oral presentation