ABSTRACT
The incidence rate of melanoma varies across regions, with Europe, the United States, and Australia having 10-25, 20-30, and 50-60 cases per 1 00 000 people. In China, patients with melanoma exhibit different clinical manifestations, pathogenesis, and outcomes. Current treatments include surgery, adjuvant therapy, and immune checkpoint inhibitors. Nonetheless, complications may arise during treatment. Melanoma development is heavily reliant on cell adhesion molecules (CAMs), and studying these molecules could provide new research directions for metastasis and progression. CAMs include the integrin, immunoglobulin, selectin, and cadherin families, and they affect multiple processes, such as maintenance, morphogenesis, and migration of adherens junction. In this study, a cell adhesion-related risk prognostic signature was constructed using bioinformatics methods, and survival analysis was performed. Plakophilin 1 (PKP1) was observed to be crucial to the immune microenvironment and has significant effects on melanoma cell proliferation, migration, invasion, and the cell cycle. This signature demonstrates high reliability and has potential for clinical applications.
Subject(s)
Melanoma , Humans , Melanoma/pathology , Cell Adhesion , Plakophilins/metabolism , Reproducibility of Results , Cadherins/metabolism , Cell Adhesion Molecules , Tumor MicroenvironmentABSTRACT
In this study, genes linked to prognosis in skin cutaneous melanoma (SKCM) involved in programmed cell death (PCD) were identified and confirmed and prognostic models based on these genes were constructed. Acquisition and analysis of clinical data and RNA sequencing information from The Cancer Genome Atlas-SKCM (TCGA-SKCM) and Sangerbox databases, gene expression data for 477 tumor samples and 2 normal samples were successfully gathered. The patients were separated into two clusters based on consensus clustering of PCD-related genes, with Cluster A having greater tumor purity, ESTIMATE score, immune score, and matrix score, and Cluster B having a significantly distinct pattern of immune cell infiltration. The use of gene set enrichment analysis and weighted correlation network analysis showed significant associations between certain genes and factors such as tumor mutation burden, age, stage, grade, and tumor subtype. Finally, based on the 12 genes selected by Least Absolute Shrinkage and Selection Operator regression analysis (STAT3, IRF2, SLC7A11, ZEB1, LIPT1, PML, GCH1, GYS1, ABCC1, XBP1, TFAP2C, NOX4), a prognostic model of PGD-related genes was constructed. The effectiveness of the model's prognostic value was confirmed through survival analysis, time-dependent receiver operating characteristic curve, single-factor Cox regression analysis, and nomogram. We also verified the relationship between the GCH1 and MKI67 expression by wet experiment. This model has high prediction accuracy in SKCM patients and can provide a reference for clinical treatment.
Subject(s)
Melanoma , Skin Neoplasms , Humans , Immunotherapy , Biomarkers , Apoptosis , Gene ExpressionABSTRACT
Changing attitudes in diplomatic relations is a common feature of international politics. However, such changes may trigger risky domino-like cascades of "friend-to-enemy" transitions among other counties and yielding catastrophic damage that could reshape the global network of international relationships. While previous attention has been focused on studying single pairs of international relationships, due to the lack of a systematic framework, it remains still unknown whether, and how, a single transition of attitude between two countries could trigger a cascade of attitude transitions among other countries. Here, we develop such a framework and construct a global evolving network of relations between country pairs based on 70,756,728 international events between 1,225 country pairs from January 1995 to March 2020. Our framework can identify and quantify the cascade of transitions following a given original transition. Surprisingly, weaker transitions are found to initiate most of the largest cascades. We also find that transitions are not only related to the balance of the local environment, but also global network properties such as betweenness centrality. Our results suggest that these transitions have a substantial impact on bilateral trade volumes and scientific collaborations. Our results reveal reaction chains of international relations, which could be helpful for designing early warning signals and mitigation methods for global international conflicts.
ABSTRACT
Fine-grained visual categorization (FGVC) relies on hierarchical features extracted by deep convolutional neural networks (CNNs) to recognize closely alike objects. Particularly, shallow layer features containing rich spatial details are vital for specifying subtle differences between objects but are usually inadequately optimized due to gradient vanishing during backpropagation. In this article, hierarchical self-distillation (HSD) is introduced to generate well-optimized CNNs features for accurate fine-grained categorization. HSD inherits from the widely applied deep supervision and implements multiple intermediate losses for reinforced gradients. Besides that, we observe that the hard (one-hot) labels adopted for intermediate supervision hurt the performance of FGVC by enforcing overstrict supervision. As a solution, HSD seeks self-distillation where soft predictions generated by deeper layers of the network are hierarchically exploited to supervise shallow parts. Moreover, self-information entropy loss (SIELoss) is designed in HSD to adaptively soften intermediate predictions and facilitate better convergence. In addition, the gradient detached fusion (GDF) module is incorporated to produce an ensemble result with multiscale features via effective feature fusion. Extensive experiments on four challenging fine-grained datasets show that, with neglectable parameter increase, the proposed HSD framework and the GDF module both bring significant performance gains over different backbones, which also achieves state-of-the-art classification performance.
ABSTRACT
Zero-shot learning (ZSL) aims at training a classification model with data only from seen categories to recognize data from disjoint unseen categories. Domain shift and generalization capability are two fundamental challenges in ZSL. In this paper, we address them with a novel Soft-Target Semi-supervised Classification (STSC) model. Specifically, an autoencoder network is leveraged, where both labeled seen data from the seen categories and unlabeled ancillary data collected from Internet or other datasets are employed as two branches, respectively. For the branch of labeled seen data, side information are employed as the latent vectors to separately connect the input of encoder and the output of decoder. In this way, visual and side information are implicitly aligned. For the branch of unlabeled ancillary data, it explicitly strengthens the reconstruction ability of the network. Meanwhile, these ancillary data can be viewed as a smooth to the domain distribution, which contributes to the alleviation of the domain shift problem. To further guarantee the generation ability, a Softmax-T loss function is proposed by making full use of the soft target. Extensive experiments on three benchmark datasets show the superiority of the proposed approach under tasks of both traditional zero-shot learning and generalized zero-shot learning.
ABSTRACT
Despite the great potential of deep neural networks (DNNs), they require massive weights and huge computational resources, creating a vast gap when deploying artificial intelligence at low-cost edge devices. Current lightweight DNNs, achieved by high-dimensional space pre-training and post-compression, present challenges when covering the resources deficit, making tiny artificial intelligence hard to be implemented. Here we report an architecture named random sketch learning, or Rosler, for computationally efficient tiny artificial intelligence. We build a universal compressing-while-training framework that directly learns a compact model and, most importantly, enables computationally efficient on-device learning. As validated on different models and datasets, it attains substantial memory reduction of ~50-90× (16-bits quantization), compared with fully connected DNNs. We demonstrate it on low-cost hardware, whereby the computation is accelerated by >180× and the energy consumption is reduced by ~10×. Our method paves the way for deploying tiny artificial intelligence in many scientific and industrial applications.
ABSTRACT
Hyperspectral imaging provides rich spatial-spectral-temporal information with wide applications. However, most of the existing hyperspectral imaging systems require light splitting/filtering devices for spectral modulation, making the system complex and expensive, and sacrifice spatial or temporal resolution. In this paper, we report an end-to-end deep learning method to reconstruct hyperspectral images directly from a raw mosaic image. It saves the separate demosaicing process required by other methods, which reconstructs the full-resolution RGB data from the raw mosaic image. This reduces computational complexity and accumulative error. Three different networks were designed based on the state-of-the-art models in literature, including the residual network, the multiscale network and the parallel-multiscale network. They were trained and tested on public hyperspectral image datasets. Benefiting from the parallel propagation and information fusion of different-resolution feature maps, the parallel-multiscale network performs best among the three networks, with the average peak signal-to-noise ratio achieving 46.83dB. The reported method can be directly integrated to boost an RGB camera for hyperspectral imaging.
ABSTRACT
Links in most real networks often change over time. Such temporality of links encodes the ordering and causality of interactions between nodes and has a profound effect on network dynamics and function. Empirical evidence has shown that the temporal nature of links in many real-world networks is not random. Nonetheless, it is challenging to predict temporal link patterns while considering the entanglement between topological and temporal link patterns. Here, we propose an entropy-rate-based framework, based on combined topological-temporal regularities, for quantifying the predictability of any temporal network. We apply our framework on various model networks, demonstrating that it indeed captures the intrinsic topological-temporal regularities whereas previous methods considered only temporal aspects. We also apply our framework on 18 real networks of different types and determine their predictability. Interestingly, we find that, for most real temporal networks, despite the greater complexity of predictability brought by the increase in dimension, the combined topological-temporal predictability is higher than the temporal predictability. Our results demonstrate the necessity for incorporating both temporal and topological aspects of networks in order to improve predictions of dynamical processes.
ABSTRACT
Radiation-induced tumor cells death is the theoretical basis of tumor radiotherapy. Death signaling disorder is the most important factor for radioresistance. However, the signaling pathway(s) leading to radiation-triggered cell death is (are) still not completely known. To better understand the cell death signaling induced by radiation, the immortalized mouse embryonic fibroblast (MEF) deficient in "initiator" caspases, "effector" caspases or different Bcl-2 family proteins together with human colon carcinoma cell HCT116 were used. Our data indicated that radiation selectively induced the activation of caspase-9 and caspase-3/7 but not caspase-8 by triggering mitochondrial outer membrane permeabilization (MOMP). Importantly, the role of radiation in MOMP is independent of the activation of both "initiator" and "effector" caspases. Furthermore, both proapoptotic and antiapoptotic Bcl-2 family proteins were involved in radiation-induced apoptotic signaling. Overall, our study indicated that radiation specifically triggered the intrinsic apoptotic signaling pathway through Bcl-2 family protein-dependent mitochondrial permeabilization, which indicates targeting mitochondria is a promising strategy for cancer radiotherapy.
Subject(s)
Apoptosis/radiation effects , Mitochondria/radiation effects , Neoplasms/genetics , Proto-Oncogene Proteins c-bcl-2/genetics , Animals , Apoptosis/genetics , Caspase 3/genetics , Caspase 7/genetics , Caspase 9/genetics , Cell Death , Fibroblasts/radiation effects , HCT116 Cells , Humans , Mice , Mitochondria/genetics , Mitochondrial Transmembrane Permeability-Driven Necrosis/radiation effects , Neoplasms/pathology , Neoplasms/radiotherapyABSTRACT
Coded-illumination (CI) imaging is a feasible technique enabling resolution enhancement and high-dimensional information extraction in optical systems. It incorporates optical encoding and computational reconstruction together to help overcome physical limitations. Existing CI reconstruction methods suffer from a trade-off between noise robustness and low computational complexity, which are both requisite for practical applications. In this paper, we propose a novel noise-robust and low-complexity reconstruction scheme for CI imaging. The scheme runs in an iterative way, and each iteration consists of two phases. First, the measurements are input into a novel non-uniform and adaptive weighted solver, whose weight updates in each iteration. This enables effective identification and attenuation of various measurement noise from coarse to fine. Second, the preserved latent information enters an alternating projection optimization procedure, which reconstructs target image by imposing support constraints without matrix lifting. We have successfully applied the scheme to structured illumination imaging and Fourier ptychography. Both simulations and experiments demonstrate that the method obtains strong robustness, low computational complexity, and fast convergence. The scheme can be adopted for various incoherent and coherent CI imaging modalities with wide extensions.
ABSTRACT
Video classification has been extensively researched in computer vision due to its wide spread applications. However, it remains an outstanding task because of the great challenges in effective spatial-temporal feature extraction and efficient classification with high-dimensional video representations. To address these challenges, in this paper, we propose an end-to-end learning framework called deep ensemble machine (DEM) for video classification. Specifically, to establish effective spatio-temporal features, we propose using two deep convolutional neural networks (CNNs), i.e., vision and graphics group and C3-D to extract heterogeneous spatial and temporal features for complementary representations. To achieve efficient classification, we propose ensemble learning based on random projections aiming to transform high-dimensional features into a set of lower dimensional compact features in subspaces; an ensemble of classifiers is trained on the subspaces and combined with a weighting layer during the backpropagation. To further enhance the performance, we introduce rectified linear encoding (RLE) inspired from error-correcting output coding to encode the initial outputs of classifiers, followed by a softmax layer to produce the final classification results. DEM combines the strengths of deep CNNs and ensemble learning, which establishes a new end-to-end learning architecture for more accurate and efficient video classification. We show the great effectiveness of DEM by extensive experiments on four data sets for diverse video classification tasks including action recognition and dynamic scene classification. Results have shown that DEM achieves high performance on all tasks with an improvement of up to 13% on CIFAR10 data set over the baseline model.
ABSTRACT
As a quorum-sensing molecule for bacteria-bacteria communication, N-(3-oxododecanoyl)-homoserine lactone (C12) has been found to possess pro-apoptotic activities in various cell culture models. However, the detailed mechanism of how this important signaling molecule function in the cells of live animals still remains largely unclear. In this study, we systematically investigated the mechanism for C12-mediated apoptosis and studied its anti-tumor effect in Caenorhabditis elegans (C. elegans). Our data demonstrated that C12 increased C. elegans germ cell apoptosis, by triggering mitochondrial outer membrane permeabilization (MOMP) and elevating the reactive oxygen species (ROS) level. Importantly, C12-induced ROS increased the expression of genes critical for DNA damage response (hus-1, clk-2 and cep-1) and genes involved in p38 and JNK/MAPK signaling pathway (nsy-1, sek-1, pmk-1, mkk-4 and jnk-1). Furthermore, C12 failed to induce germ cell apoptosis in animals lacking the expression of each of those genes. Finally, in a C. elegans tumor-like symptom model, C12 significantly suppressed tumor growth through inhibiting the expression of RAS/MAPK pathway genes (let-23/EGFR, let-60/RAS, lin-45/RAF, mek-2/MEK and mpk-1/MAPK). Overall, our results indicate that DNA damage response and MAPK activation triggered by mitochondrial ROS play important roles in C12-induced apoptotic signaling in C. elegans, and RAS/MAPK suppression is involved in the tumor inhibition effect of C12. This study provides in vivo evidence that C12 is a potential candidate for cancer therapeutics by exerting its pro-apoptotic and anti-tumor effects via elevating mitochondria-dependent ROS production.
Subject(s)
4-Butyrolactone/analogs & derivatives , Apoptosis/drug effects , Carcinogenesis/drug effects , Germ Cells/pathology , MAP Kinase Signaling System/drug effects , Mitochondria/drug effects , Reactive Oxygen Species/metabolism , 4-Butyrolactone/pharmacology , Animals , Apoptosis/genetics , Caenorhabditis elegans/drug effects , Caenorhabditis elegans/genetics , Caenorhabditis elegans/metabolism , Caenorhabditis elegans Proteins/genetics , Caenorhabditis elegans Proteins/metabolism , Carcinogenesis/metabolism , DNA Damage/drug effects , DNA Damage/genetics , Female , Germ Cells/drug effects , MAP Kinase Signaling System/genetics , Mitochondria/metabolism , Mutation , Oxidative Stress , RNA Interference , ras Proteins/genetics , ras Proteins/metabolismABSTRACT
Despite treatment of lung cancer with radiotherapy and chemotherapy, the survival rate of lung cancer patients remains poor. Previous studies demonstrated the importance of upregulation of inflammatory factors, such as cyclooxygenase 2 (cox2), in tumor tolerance. In the present study, we investigated the role of cox2 in radiosensitivity of lung cancer. Our results showed that the combination treatment of radiation with aspirin, an anti-inflammatory drug, induced a synergistic reduction of cell survival in A549 and H1299 lung cancer cells. In comparison with normal human lung fibroblasts (NHLFs), the cell viability was significantly decreased and the level of apoptosis was remarkably enhanced in A549 cells. Mechanistic studies revealed that the reduction of cox2 by aspirin in A549 and H1299 was caused by disruption of the chromosomal architecture of the cox2 locus. Moreover, the disruption of chromatin looping was mediated by the inhibition of nuclear translocation of p65 and decreased enrichment of p65 at cox2-regulatory elements. Importantly, disorganization of the chromosomal architecture of cox2 triggered A549 cells sensitive to γ-radiation by the induction of apoptosis. In conclusion, we present evidence of an effective therapeutic treatment targeting the epigenetic regulation of lung cancer and a potential strategy to overcome radiation resistance in cancer cells.
Subject(s)
Cyclooxygenase 2/genetics , Genetic Therapy , Lung Neoplasms/radiotherapy , Radiation Tolerance/drug effects , A549 Cells , Apoptosis/drug effects , Aspirin/pharmacology , Cell Proliferation/drug effects , Cell Survival/drug effects , Chromatin/genetics , Combined Modality Therapy , Cyclooxygenase 2 Inhibitors/pharmacology , Epigenesis, Genetic/genetics , Humans , Lung/drug effects , Lung/pathology , Lung Neoplasms/drug therapy , Lung Neoplasms/genetics , Signal Transduction/drug effects , eIF-2 Kinase/geneticsABSTRACT
Radiotherapy is used to treat approximately 50% of all cancer patients, with varying prognoses. Intrinsic radiosensitivity is an important factor underlying the radiotherapeutic efficacy of this precise treatment. During the past decades, great efforts have been made to improve radiotherapy treatment through multiple strategies. However, invaluable data remains buried in the extensive radiotherapy literature, making it difficult to obtain an overall view of the detailed mechanisms leading to radiosensitivity, thus limiting advances in radiotherapy. To address this issue, we collected data from the relevant literature contained in the PubMed database and developed a literature-based database that we term the cancer radiosensitivity regulation factors database (dbCRSR). dbCRSR is a manually curated catalogue of radiosensitivity, containing multiple radiosensitivity regulation factors (395 coding genes, 119 non-coding RNAs and 306 chemical compounds) with appropriate annotation. To illustrate the value of the data we collected, data mining was performed including functional annotation and network analysis. In summary, dbCRSR is the first literature-based database to focus on radiosensitivity and provides a resource to better understand the detailed mechanisms of radiosensitivity. We anticipate dbCRSR will be a useful resource to enrich our knowledge and to promote further study of radiosensitivity.Database URL: http://bioinfo.ahu.edu.cn: 8080/dbCRSR/.
Subject(s)
Data Curation , Data Mining , Databases, Bibliographic , Neoplasms/metabolism , Radiation Tolerance , Animals , Humans , Neoplasms/pathology , Neoplasms/radiotherapyABSTRACT
In this paper, we propose YoTube-a novel deep learning framework for generating action proposals in untrimmed videos, where each action proposal corresponds to a spatial-temporal tube that potentially locates one human action. Most of the existing works generate proposals by clustering low-level features or linking image proposals, which ignore the interplay between long-term temporal context and short-term cues. Different from these works, our method considers the interplay by designing a new recurrent YoTube detector and static YoTube detector. The recurrent YoTube detector sequentially regresses candidate bounding boxes using Recurrent Neural Network learned long-term temporal contexts. The static YoTube detector produces bounding boxes using rich appearance cues in every single frame. To fully exploit the complementary appearance, motion, and temporal context, we train the recurrent and static detector using RGB (Color) and flow information. Moreover, we fuse the corresponding outputs of the detectors to produce accurate and robust proposal boxes and obtain the final action proposals by linking the proposal boxes using dynamic programming with a novel path trimming method. Benefiting from the pipeline of our method, the untrimmed video could be effectively and efficiently handled. Extensive experiments on the challenging UCF-101, UCF-Sports, and JHMDB datasets show superior performance of the proposed method compared with the state of the arts.
ABSTRACT
Visual target tracking is one of the most important research areas in the field of computer vision. Within this realm, multiple targets tracking (MTT) under complicated scene stands out for its great availability in real life applications, such as urban traffic surveillance and sports video analysis. However, in MTT, main difficulties arise from large variation in target saliency and significant motion heterogeneity, which may result in the failure of tracking weak targets. To tackle this challenge, a novel hierarchical layered tracking structure is proposed to perform tracking sequentially layer-by-layer. Upon this layered structure, we establish an intertarget mutual assistance mechanism on basis of intertarget correlation exploited among targets. The tracking results of a subset of targets can be utilized as additional prior information for tracking other targets. Specifically, a nonlinear motion model as well as a target interaction model basing on the intertarget correlation are proposed to effectively estimate the possible target region-of-interest to facilitate the prediction-based tracking. Moreover, the concept of motion entropy is introduced to quantitatively measure the degree of motion heterogeneity within the tracking scene for layer construction. Compared to other existing methods, extensive experiments demonstrated that the proposed method is capable of achieving higher tracking performance in complicated scenes, where targets are characterized with great heterogeneity.
ABSTRACT
In the 1980s, Global Positioning System (GPS) receiver autonomous integrity monitoring (RAIM) was proposed to provide the integrity of a navigation system by checking the consistency of GPS measurements. However, during the approach and landing phase of a flight path, where there is often low GPS visibility conditions, the performance of the existing RAIM method may not meet the stringent aviation requirements for availability and integrity due to insufficient observations. To solve this problem, a new RAIM method, named vision-aided RAIM (VA-RAIM), is proposed for GPS integrity monitoring in the approach and landing phase. By introducing landmarks as pseudo-satellites, the VA-RAIM enriches the navigation observations to improve the performance of RAIM. In the method, a computer vision system photographs and matches these landmarks to obtain additional measurements for navigation. Nevertheless, the challenging issue is that such additional measurements may suffer from vision errors. To ensure the reliability of the vision measurements, a GPS-based calibration algorithm is presented to reduce the time-invariant part of the vision errors. Then, the calibrated vision measurements are integrated with the GPS observations for integrity monitoring. Simulation results show that the VA-RAIM outperforms the conventional RAIM with a higher level of availability and fault detection rate.
Subject(s)
Geographic Information Systems , Algorithms , Aviation/methods , Computer SimulationABSTRACT
In this paper, we study the problem of detecting sudden pedestrian crossings to assist drivers in avoiding accidents. This application has two major requirements: to detect crossing pedestrians as early as possible just as they enter the view of the car-mounted camera and to maintain a false alarm rate as low as possible for practical purposes. Although many current sliding-window-based approaches using various features and classification algorithms have been proposed for image-/video-based pedestrian detection, their performance in terms of accuracy and processing speed falls far short of practical application requirements. To address this problem, we propose a three-level coarse-to-fine video-based framework that detects partially visible pedestrians just as they enter the camera view, with low false alarm rate and high speed. The framework is tested on a new collection of high-resolution videos captured from a moving vehicle and yields a performance better than that of state-of-the-art pedestrian detection while running at a frame rate of 55 fps.
Subject(s)
Accidents, Traffic/prevention & control , Algorithms , Artificial Intelligence , Automobile Driving , Decision Support Techniques , Models, Theoretical , Pattern Recognition, Automated/methods , Computer SimulationABSTRACT
Classification-based pedestrian detection systems (PDSs) are currently a hot research topic in the field of intelligent transportation. A PDS detects pedestrians in real time on moving vehicles. A practical PDS demands not only high detection accuracy but also high detection speed. However, most of the existing classification-based approaches mainly seek for high detection accuracy, while the detection speed is not purposely optimized for practical application. At the same time, the performance, particularly the speed, is primarily tuned based on experiments without theoretical foundations, leading to a long training procedure. This paper starts with measuring and optimizing detection speed, and then a practical classification-based pedestrian detection solution with high detection speed and training speed is described. First, an extended classification/detection speed metric, named feature-per-object (fpo), is proposed to measure the detection speed independently from execution. Then, an fpo minimization model with accuracy constraints is formulated based on a tree classifier ensemble, where the minimum fpo can guarantee the highest detection speed. Finally, the minimization problem is solved efficiently by using nonlinear fitting based on radial basis function neural networks. In addition, the optimal solution is directly used to instruct classifier training; thus, the training speed could be accelerated greatly. Therefore, a rapid and accurate classification-based detection technique is proposed for the PDS. Experimental results on urban traffic videos show that the proposed method has a high detection speed with an acceptable detection rate and a false-alarm rate for onboard detection; moreover, the training procedure is also very fast.
Subject(s)
Biometry/methods , Image Processing, Computer-Assisted/methods , Neural Networks, Computer , Walking/physiology , Accidents, Traffic/prevention & control , Algorithms , Automobiles , HumansABSTRACT
With the rapid development of the economy and the accelerated globalization process, the aviation industry plays a more and more critical role in today's world, in both developed and developing countries. As the infrastructure of aviation industry, the airport network is one of the most important indicators of economic growth. In this paper, we investigate the evolution of the Chinese airport network (CAN) via complex network theory. It is found that although the topology of CAN has remained steady during the past few years, there are many dynamic switchings inside the network, which have changed the relative importance of airports and airlines. Moreover, we investigate the evolution of traffic flow (passengers and cargoes) on CAN. It is found that the traffic continues to grow in an exponential form and has evident seasonal fluctuations. We also found that cargo traffic and passenger traffic are positively related but the correlations are quite different for different kinds of cities.