Supervision Information Research Articles

Complementary label learning based on knowledge distillation.

Complementary label learning (CLL) is a type of weakly supervised learning method that utilizes the category of samples that do not belong to a certain class to learn their true category. However, current CLL methods mainly rely on rewriting classification losses without fully leveraging the supervisory information in complementary labels. Therefore, enhancing the supervised information in complementary labels is a promising approach to improve the performance of CLL. In this paper, we propose a novel framework called Complementary Label Enhancement based on Knowledge Distillation (KDCL) to address the lack of attention given to complementary labels. KDCL consists of two deep neural networks: a teacher model and a student model. The teacher model focuses on softening complementary labels to enrich the supervision information in them, while the student model learns from the complementary labels that have been softened by the teacher model. Both the teacher and student models are trained on the dataset that contains only complementary labels. To evaluate the effectiveness of KDCL, we conducted experiments on four datasets, namely MNIST, F-MNIST, K-MNIST and CIFAR-10, using two sets of teacher-student models (Lenet-5+MLP and DenseNet-121+ResNet-18) and three CLL algorithms (PC, FWD and SCL-NL). Our experimental results demonstrate that models optimized by KDCL outperform those trained only with complementary labels in terms of accuracy.

Slow Motion Matters: A Slow Motion Enhanced Network for Weakly Supervised Temporal Action Localization

Weakly supervised temporal action localization (WTAL) aims to localize actions in untrimmed videos with only weak supervision information (e.g. video-level labels). Most existing models handle all input videos with a fixed temporal scale. However, such models are not sensitive to actions whose pace of the movements is different from the ``normal" speed, especially slow-motion action instances, which complete the movements with a much slower speed than their counterparts with a normal speed. Here arises the slow-motion blurred issue: It is hard to explore salient slow-motion information from videos at ``normal" speed. In this paper, we propose a novel framework termed Slow Motion Enhanced Network (SMEN) to improve the ability of a WTAL network by compensating its sensitivity on slow-motion action segments. The proposed SMEN comprises a Mining module and a Localization module. The mining module generates mask to mine slow-motion-related features by utilizing the relationships between the normal motion and slow motion; while the localization module leverages the mined slow-motion features as complementary information to improve the temporal action localization results. Our proposed framework can be easily adapted by existing WTAL networks and enable them be more sensitive to slow-motion actions. Extensive experiments on three benchmarks are conducted, which demonstrate the high performance of our proposed framework.

Underwater Forward-Looking Sonar Images Target Detection via Speckle Reduction and Scene Prior

Forward-looking sonar (FLS) imagery system plays a significant role in oceanic object recognition and detection since it can overcome the limitation of lighting conditions and reflect the real situation of the underwater environment. However, object detection algorithms for FLS images remain challenging for two main reasons: 1) The noise caused by the coherent characteristic of the scattering phenomenon impairs the detector capture of target information; 2) The scene prior based on the uneven target scale distribution is generally neglected, which leads to the detector generating redundant anchors and slows down detection efficiency. Confronting such challenges, this paper characterizes the noise and the uneven target scale distribution in FLS images as multiplicative speckle noise and scene prior, respectively. Therefore, we propose a novel underwater FLS image detection network, namely UFIDNet, to further improve detection performance by considering speckle noise reduction and scene prior in FLS images. More specifically, a speckle reduction auxiliary branch (SRAB) is designed to introduce additional despeckled supervision information to encourage the feature extractor to produce clean features and share them with the detection pipeline during the training phase. In particular, the noise distribution of FLS images is excavated for synthetic dataset construction and Despeckle Network (DSN) design to obtain despeckled supervision images. In addition, a feature selection strategy (FSS) embedded in detection branch is designed to screen out feature levels that do not match the target size, thus significantly reducing the generation of redundant anchors and improving detection speed. Experimental results show that our UFIDNet achieves 70.5% and 47.3% average precision (AP), 81.3% and 54.6% average recall (AR <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">max=10</sub> ), 27.0 and 26.1 FPS on two real FLS datasets respectively, outperforming many state-of-the-art general detectors and sonar image detectors.

Semi-supervised method for improving general-purpose and domain-specific textual corpora labels

The study is devoted to the problem of label quality assurance and labeling cost minimization. It is important to have high quality labels in order to build efficient supervised machine learning pipelines. The aforementioned labels have proved to be costly to acquire; however, one can work with weaker instances of supervision that may be helpful to create a reliably set of ground truth labels for classifiers. The authors present a semi-supervised approach using pretrained language models to increase supervision information quality from general-purpose as well domain-specific document sets where weak labels and unlabeled data is present.

Sistem Infomasi Pengawasan Sekolah Dasar dan Menengah di Dinas Pendidikan Pemuda dan Olahraga Menggunakan Metode Waterfall

The importance of a computerized system will save more time in managing ongoing business processes and produce accuracy in presenting data or reports to government institutions. The Cianjur Education, Youth, Sports Office requires a data management system for primary and secondary school supervision, namely an Information System that can be used by supervisors to take an inventory of data, including recording, storing, and reporting data from primary and secondary schools. With the information system, the Education, Youth, and Sports Office of Cianjur Regency can see information from schools quickly, precisely, and accurately. Data collection techniques in this study include observation, interviews and documentation. While the research method used is the waterfall method with stages, Communication, Planning, Modeling, Construction, Deployment, Testing. The results of this study are in the form of an analysis design for primary and secondary school supervision information systems using UML design tools, as well as the results of testing the system with the EUCS method.

Weakly-Supervised Semantic Segmentation Network With Iterative dCRF

This Autonomous driving methods driven by big data are becoming more and more perfect, but the cost of existing data labeling is too high, so how to reduce or even not label data has attracted more and more attention. Semantic segmentation networks supervised by image-level annotations are all trained using pseudo-labels. Most methods use image classification networks to generate class activation maps (CAMs) and start with CAMs to diffuse features to other parts of the target to obtain pseudo-labels. However, due to its weak supervision information, it is difficult for the existing methods to obtain better results. Therefore, we propose a weakly-supervised semantic segmentation network with iterative dCRF based on graph convolution. Specifically, we use ResNet to generate CAMs and node features and then use graph convolution for feature propagation and merge the low-level and high-level semantic information of the image. Then execute dCRF in an iterative manner, and finally obtain refined pseudo-labels. On the PASCAL VOC 2012 data set, our model achieves an mIoU of 63.5%, which is 0.3% higher than the graph convolutional network method.

Time Series Artificial Neural Network based Classification Model for Asset Trading Strategy

Best daily trading strategy is meaningful to investors to obtain the maximum return. Financial market plays a key role in the social development. Sufficient funds are of great significance to promote the development of all sectors of society. From the perspective of investors, their purpose is to obtain the maximum income through investing assets in the market. At the same time, the benefit of investors can attract more investors and funds. However, there are huge risks in investment in financial markets. This is mainly because the changes of assets' prices are very complex, and traders cannot predict the changes of assets' prices. Therefore, it is of great significance for investors to determine the trading behavior according to the historical data of asset's price. Aiming at the problem of asset investment strategy, we study from three aspects: asset value accounting, optimal investment strategy and investment decision prediction. A novel method is proposed to represent trading action based on asset state, which represents the current asset state through a 0-1 variable. Then, the change of state indicates the occurrence of transaction. Based on this variable, we propose an evaluation model of the total value of the asset on every day. By maximizing the value of the combination, we establish an optimization model and the Genetic Algorithm (GA) is used to search the best investment combination. Using the prior information of the price, we first solve the combination of asset states of the investor corresponding to the best trading strategy. Then, taking the sequence of states as the supervision information, an Artificial Neural Network (ANN) based classification model for asset state prediction is established. The historical data of Bitcoin daily price is used to predict the asset state on the current day, to realize the decision-making of daily trading strategy. Extensive experiments have been conducted, and the experimental results verify our model can effectively find potential pattern for asset investment and achieve good profits based on the predict investment decisions.

Parallel learned generative adversarial network with multi-path subspaces for cross-modal retrieval

Cross-modal retrieval aims to narrow the heterogeneity gap between different modalities, such as retrieving images through texts or vice versa. One of the key challenges of cross-modal retrieval is the inconsistent distribution across diverse modalities. Most existing methods tend to construct a common representation subspace to overcome the challenge. However, the supervision information is not fully explored in most single-path cross-modal learning approaches. In this paper, we present a novel Parallel Learned generative adversarial network with Multi-path Subspaces (PLMS) for cross-modal retrieval. PLMS is a parallel learned architecture that aims to capture more effective information in an end-to-end trained cross-modal retrieval model. To be specific, a dual-branch network is constructed in the modality-specific generator, thereby the overall framework learns two common subspaces to emphasize discrepant supervision information and preserve more effective transformed features. We further design two objective functions for the training of the dual branches in generators. Through joint training, the feature representations generated by dual branches in a specific modality are fused for similarity measurement between modalities. To avoid redundancy and overlap during fusion, a Multi-source Domain Balancing (MDB) mechanism is presented to explore the contribution of the two specific-task branches. Extensive experiments show that our proposed method is effective and achieves state-of-the-art results on four widely-used databases.

SMGCL: Semi-supervised Multi-view Graph Contrastive Learning

Graph contrastive learning (GCL), aiming to generate supervision information by transforming the graph data itself, is increasingly becoming a focus of graph research. It has shown promising performance in graph representation learning by extracting global-level abstract features of graphs. Nonetheless, most GCL methods are performed in a completely unsupervised manner and would get unappealing results in balancing the multi-view information of graphs. To alleviate this, we propose a Semi-supervised Multi-view Graph Contrastive Learning (SMGCL) framework for graph classification. The framework can capture the comparative relations between label-independent and label-dependent node (or graph) pairs across different views. In particular, we devise a graph neural network (GNN)-based label augmentation module to exploit the label information and guarantee the discrimination of the learned representations. In addition, a shared decoder module is complemented to extract the underlying determinative relationship between learned representations and graph topology. Experimental results on graph classification tasks demonstrate the superiority of the proposed framework.

Anchor-free temporal action localization via Progressive Boundary-aware Boosting

Enormous untrimmed videos from the real world are difficult to analyze and manage. Temporal action localization algorithms can help us to locate and recognize human activity clips in untrimmed videos. Recently, anchor-free temporal action localization methods have gained increasing attention due to small computational costs and no complex hyperparameters of pre-set anchors. Although the performance has been significantly improved, most existing anchor-free temporal action localization methods still suffer from inaccurate action boundary predictions. In this paper, we want to alleviate the above problem through boundary refinement and temporal context aggregation. To this end, a novel Progressive Boundary-aware Boosting Network (PBBNet) is proposed for anchor-free temporal action localization. The PBBNet consists of three main modules: Temporal Context-aware Module (TCM), Instance-wise Boundary-aware Module (IBM), and Frame-wise Progressive Boundary-aware Module (FPBM). The TCM aggregates the temporal context information and provides features for the IBM and the FPBM. The IBM generates multi-scale video features to predict action results coarsely. Compared with IBM, the FPBM focuses on instance features corresponding to action predictions and uses more supervision information for boundary regression. Given action results from IBM, the FPBM uses a progressive boosting strategy to refine the boundary predictions multiple times with supervision from weak to strong. Extensive experiments on three benchmark datasets THUMOS14, ActivityNet-v1.3 and HACS show our PPBNet outperforms all existing anchor-free methods. Further, our PPBNet achieves state-of-the-art performance (72.5% mAP at tIoU = 0.5) on THUMOS14 dataset.

Attention-Aggregated Attribute-Aware Network With Redundancy Reduction Convolution for Video-Based Industrial Smoke Emission Recognition

Existing video-based industrial smoke emission recognition methods face the issues of low detection rates and high false alarm rates. An important reason is that they only consider binary category information as supervision information and ignore video attribute information, which provides important supplementary in improving model performances. To solve it, we propose an attention-aggregated attribute-aware network (AANet). First, to effectively guide the model for discriminative feature learning, a video attribute information decoding module is proposed to increase supervision information by designing attribute vector construction and attribute information decoding methods. Second, to learn discriminative feature representations, some attentions are designed to aggregate spatiotemporal and context information based on ConvLSTM, global feature extraction, and cascaded pyramid attention. Final, the redundancy reduction convolution is proposed to reduce redundant channels by channelwise weights considering matrix elements summation and information spatial distribution characterized by information entropy. Extensive experiments show that AANet significantly outperforms existing methods

Research and Practice on Project based Teaching of Java Web Application Development Course based on Shipping Characteristics

To overcoming the separation problem between the major teaching and industrial development, the software technology specialty takes the systematization of the working process as the basic idea, carries out the project-based teaching reform, and realizes the transformation from the knowledge system to the ability system. In order to better reflect the shipping characteristics, the software technology specialty takes the shipping information project as a project-based teaching case. As a core course, the Java Web application development course takes the enterprise workbench module of ship water pollution supervision and service information system as a teaching case throughout the course. In the teaching process, the cultivation of knowledge, skills and accomplishments is carried out throughout the process of student project realization, so as to truly achieve learning by doing and learning by doing. The project teaching of Java Web application development course with shipping characteristics has improved the teaching quality of teachers and students' enthusiasm, and provided a new method and perspective for the teaching reform of software technology courses in higher vocational colleges.

Distributed Supervision Model for Enterprise Data Asset Trading Based on Blockchain Multi-Channel in Industry Alliance.

Compared with traditional physical commodities, data are intangible and easy to leak, and the related trading process has problems, such as complex participating roles, lengthy information flow, poor supervisory coverage and difficult information traceability. To handle these problems, we construct a distributed supervision model for data trading based on blockchain, and conduct multi-party hierarchical and multi-dimensional supervision of the whole process of data trading through collaborative supervision before the event, at present and after the event. First, the characteristics of information flow in the data trading process are analyzed, and the main subject and key supervision information in the data trading process are sorted out and refined. Secondly, combined with the actual business process of data trading supervision, a multi-channel structure of distributed supervision is proposed by adopting an access–verification–traceability strategy. Finally, under the logical framework of the supervision model, the on-chain hierarchical structure and the data hybrid storage method of “on-chain + off-chain” are designed, and multi-supervisor-oriented hierarchical supervision and post-event traceability are realized through smart contracts. The results show that the constructed blockchain-based distributed supervision model of data trading can effectively isolate and protect sensitive and private information between data trading, so as to realize the whole process, multi-subject and differentiated supervision of key information of data trading, and provide an effective and feasible method for the controllable and safe supervision of data trading.

Analysis of epidemiological characteristics of bacillary dysentery with multiple-onset in Henan Province from 2005 to 2020

Objective: To understand the epidemiological characteristics of bacillary dysentery with multiple-onset in Henan province from 2005 to 2020. Methods: The reported cases of bacillary dysentery (including confirmed cases and clinically diagnosed cases) in Henan Province from January 2005 to December 2020 were collected through China's National Disease Supervision Information Management System. The main information included gender, age, home address, date of onset and date of diagnosis. The interval between two episodes of the same case was more than 15 days, which was judged as two episodes. The incidence characteristics of bacillary dysentery patients with two or more cases in Henan Province from 2005 to 2020 were analyzed, and the regional distribution map of cases was drawn using ArcGIS software. Results: From 2005 to 2020, a total of 250 430 cases of bacillary dysentery were reported in Henan Province, with a cumulative incidence rate of 228.66/100 000. There were 2 342 cases with two or more attacks. The incidence of recurrent cases of bacillary dysentery increased year by year (χ2trend=2 932.28, P<0.001). There was no significant difference in the incidence of two or more cases of different sexes (χ2=0.39, P=0.540). There was significant difference in the incidence among different age groups (χ2=438.40, P<0.001). The incidence of two or more cases in the 60-69 age group was relatively high (1.70%). The shortest time interval between the onset of the disease was 16 days, and the longest was 5 579 days, with M (Q1, Q3) about 428 (237, 843) days. Compared with healthy people, those with a history of bacterial diseases had a higher risk of developing bacillary dysentery (RR: 4.12, 95%CI: 3.95‒4.29). Conclusion: The proportion of patients with multiple-onset shows an increasing trend, and there is an age difference.

FBCU-Net: A fine-grained context modeling network using boundary semantic features for medical image segmentation

The performance of deep learning-based medical image segmentation methods largely depends on the segmentation accuracy of tissue boundaries. However, since the boundary region is at the junction of areas of different categories, the pixels located at the boundary inevitably carry features belonging to other classes and difficult to distinguish. This paper proposes a fine-grained contextual modeling network for medical image segmentation based on boundary semantic features, FBCU-Net, which uses the semantic features of boundary regions to reduce the influence of irrelevant features on boundary pixels. First, based on the discovery that indistinguishable pixels are usually boundary pixels in medical images, we introduce new supervision information to find and classify boundary pixels. Second, based on the existing relational context modeling schemes, we generate the boundary region representations representing the semantic features of boundary regions. Last, we use boundary region representations to reduce the influence of irrelevant features on boundary pixels and generate highly discriminative pixel representations. Furthermore, to enhance the attention of the network to the boundary region, we also propose the boundary enhancement strategy. We evaluate the proposed model on five datasets, TUI (Thyroid Tumor), ISIC-2018 (Dermoscopy), 2018 Data Science Bowl (Cell Nuclei), Glas (Colon Cancer), and BUSI (Breast Cancer). The results show that FBCU-Net has better boundary segmentation performance and overall performance for different medical images than other state-of-the-art (SOTA) methods, and has great potential for clinical application.

CENet: A Channel-Enhanced Spatiotemporal Network With Sufficient Supervision Information for Recognizing Industrial Smoke Emissions

Vision-based industrial smoke emission recognition technology can identify smoke emissions and provide a visual evidence for humans to pursue environmental justice. However, the existing methods still face the issues of low detection rates (DRs) and high false alarm rates (FARs) due to the insufficient supervision information and limited feature representation capability. To solve these issues, this article presents a channel-enhanced spatiotemporal network (CENet) with sufficient supervision information for recognizing industrial smoke emissions. First, to provide sufficient supervision information for learning discriminative feature representation, we propose a new loss function by collaboratively taking the binary category, pixel-level smoke density, and background information as supervision information and use them in the network final layer as well as the network middle layers to guide model training. Second, to solve the deficiencies of max/average pooling and convolution operations in feature extraction, we propose channel-enhanced modules, including channel-enhanced pooling (CEPool), channel-enhanced convolution (CEConv), and channel-enhanced upsampling (CEUpsample) to learn high-response values of bright features as well as the low-response values with discriminative features. The channel-enhanced modules selectively enhance the learned features with large amount of information and suppress those useless features. Third, we propose a two-stage method based on spatiotemporal information extraction module (SIEM) and smoke recognition module (SRM), which are designed to learn spatiotemporal information between input frames and that between smoke density frames, respectively. Extensive experiments show that CENet achieves the best performance among the existing smoke recognition methods.

Evolutionary game model of construction enterprises and construction material manufacturers in the construction and demolition waste resource utilization.

With the continuous advancement of urbanization, a huge amount of construction and demolition waste (CDW) is generated in large-scaled construction activities, which has aggravated the problem of environmental pollution, waste of resources and destruction of city appearance. In the context of waste-free city, the recycling of CDW can reduce environmental pollution and promote the sustainable development of a city. However, only 20-30% of CDW in the world is recycled, showing a low rate of global CDW utilization. In order to improve the utilization rate, this paper selects construction enterprises and construction material manufacturers as main participants, applies evolutionary game theory to construct an evolutionary game model on the two parties' decision-making behaviors in CDW recycling, and uses MATLAB to make a numerical simulation. The aim of the model is to analyze the influence of various factors on the parties' decision-making behavior evolution and propose strategies to promote CDW utilization. The study found that the stable state of the CDW resource utilization system mainly depends on the difference between revenue and costs, the initial strategy, and the strength of the external environment; for the government, a supervision strategy is found to be necessary, and the best supervision level is 0.6. In the early stage of resource utilization of CDW, subsidies to construction material manufacturers should be increased to improve their initial participation; public participation can effectively improve the efficiency of government supervision, and its optimal participation level is greater than or equal to 0.4; under weak supervision, government penalty increases alone cannot prevent construction enterprises from illegally disposing of CDW. Therefore, the greater the difference, the positive the initial strategy, and the stronger the external environment, the more the behavior of the two participants tends to be {participation, use}. The results show that the government should establish effective supervision mechanisms and legal systems, improve supervision hotlines and information platforms, encourage the public to participate in CDW management and supervision, set appropriate rewards and punishments, strengthen supervision and management levels, reduce supervision costs, and ensure the effectiveness of construction management to improve the efficiency of cooperation between construction enterprises and construction material manufacturers.

A label-guided weighted semi-supervised neutrosophic clustering algorithm

The traditional neutrosophic clustering method only performs cluster analysis on the data itself, and often ignores the supervision information of data. In order to solve the above problems, a label-guided weighted semi-supervised neutrosophic clustering algorithm is proposed in the paper. On the one hand, the paired constraint information is used to construct the supervision weight coefficient and the distance measurement learning is combined to re-measure the degree of membership of the data and the cluster center; On the other hand, by minimizing the sum of squares of error between membership matrix and label matrix, the purpose of clustering results guided by label information is realized. Experiments on various data sets and comparisons with other clustering algorithms show that the new clustering algorithm can make full use of supervisory information and improve the accuracy of clustering.

Ensemble learning via supervision augmentation for white matter hyperintensity segmentation.

Since the ambiguous boundary of the lesion and inter-observer variability, white matter hyperintensity segmentation annotations are inherently noisy and uncertain. On the other hand, the high capacity of deep neural networks (DNN) enables them to overfit labels with noise and uncertainty, which may lead to biased models with weak generalization ability. This challenge has been addressed by leveraging multiple annotations per image. However, multiple annotations are often not available in a real-world scenario. To mitigate the issue, this paper proposes a supervision augmentation method (SA) and combines it with ensemble learning (SA-EN) to improve the generalization ability of the model. SA can obtain diverse supervision information by estimating the uncertainty of annotation in a real-world scenario that per image have only one ambiguous annotation. Then different base learners in EN are trained with diverse supervision information. The experimental results on two white matter hyperintensity segmentation datasets demonstrate that SA-EN gets the optimal accuracy compared with other state-of-the-art ensemble methods. SA-EN is more effective on small datasets, which is more suitable for medical image segmentation with few annotations. A quantitative study is presented to show the effect of ensemble size and the effectiveness of the ensemble model. Furthermore, SA-EN can capture two types of uncertainty, aleatoric uncertainty modeled in SA and epistemic uncertainty modeled in EN.

Self-supervised learning for modal transfer of brain imaging.

Today's brain imaging modality migration techniques are transformed from one modality data in one domain to another. In the specific clinical diagnosis, multiple modal data can be obtained in the same scanning field, and it is more beneficial to synthesize missing modal data by using the diversity characteristics of multiple modal data. Therefore, we introduce a self-supervised learning cycle-consistent generative adversarial network (BSL-GAN) for brain imaging modality transfer. The framework constructs multi-branch input, which enables the framework to learn the diversity characteristics of multimodal data. In addition, their supervision information is mined from large-scale unsupervised data by establishing auxiliary tasks, and the network is trained by constructing supervision information, which not only ensures the similarity between the input and output of modal images, but can also learn valuable representations for downstream tasks.