最新刊期
卷 69 ， 期 1 ， 2023
- 摘要：In order to apply blockchain transactions to daily payment scenarios and reduce the use threshold of blockchain transactions, this paper proposes a scheme based on merchant alliance for fast blockchain transactions. The scheme fully considers the actual scene requirements of daily payment and separates payment from transaction, so as to shorten the delay of payment confirmation. In view of the shortcomings in locked funds and transaction fee in the existing work, on the one hand, we design a chip trading mechanism to release the locked funds in the system and to reduce the threshold for the use of blockchain in daily payment scenarios. On the other hand，we separate merchant settlement from consumer settlement to reduce the transaction fee by designing transaction settlement mechanism. Considering the possible evil behavior in the transaction process, we provide security guarantee for the fast blockchain transaction by designing the state synchronization mechanism. Comparing the scheme with other work schemes, the results show that the scheme requires less handling fee and locked funds，which is more suitable for daily payment scenarios.关键词：blockchain transaction;daily payment;transaction fee501|231|0更新时间：2023-02-15
- 摘要：Existing studies on backdoor detection mainly focus on white-box settings. However, internal information is not always exposed to the security auditors. To bridge this gap, we investigated the black-box backdoor detection based on Monte-Carlo gradient estimation. We propose a black-box trigger reversing method to determine whether a model is infected with a backdoor attack by regarding the black-box trigger reversing problem as a zeroth-order optimization. Furthermore, we propose a fast black-box backdoor detection method to reduce the detection overhead by using the importance sampling technique, a regularized criterion, and an early stopping strategy. Experimental results on three popular image datasets show that the proposed method can accurately distinguish benign and infected models and obtain effective backdoor triggers.关键词：AI security;deep learning;neural network;neural network backdoor;security audit439|668|0更新时间：2023-02-15
- 摘要：In the transfer learning scenario with image classification as the target task, to solve the problem of low accuracy in the case of inconsistent corresponding models of the attacker's ability and demand, a membership inference attack scheme for feature vector difference is proposed, shadow model is constructed to help adversaries obtain feature vectors at different levels, and Euclidean distance is used to calculate the distance between different feature vectors. The threshold comparison step is proposed to classify Euclidean distance and a threshold selection strategy is also designed. Experimental results show that the proposed attack scheme can obtain great attack performance when attacking the teacher model without visiting the teacher model. The membership inference attack accuracy on Cats vs Dogs, Flowers102 and CIFAR-100 datasets achieves 0.728, 0.632 and 0.581 respectively, which reveals the effectiveness of membership inference attack in the transfer learning scenario. Moreover, with the increase of the number of frozen layers in student model training process, the attack performance of the proposed scheme on Cats vs Dogs datasets is improved.关键词：machine learning;membership inference attack;transfer learning217|438|0更新时间：2023-02-15
- 摘要：Privacy has been widely studied in the blockchain projects. However,the transactions information is stored in plaintext in the blockchain for public verification, which will leak relevant transaction information and cannot protect the privacy of users. The possibility of privacy leakage may severely hinder the implementation of blockchain in practice. Hence, how to securely and effectively protect the privacy of transactions is worth further research. In this paper, we proposed a homomorphic encryption algorithm PailGamal based on the analysis of decryption of Paillier algorithm and analysis of ElGamal algorithm. PailGamal supports both the direct decryption of ciphertext and the regulation of all transaction information by the regulator.Then we combine the PailGamal with the range proof scheme Bulletproofs to design a regulatable and efficient confidential transaction scheme. In this scheme, PailGamal is used to encrypt the transaction amount, and no one else can know the transaction amount except for the regulator and transaction participants; combined with the range proof, the data availability and legality are guaranteed while protecting user privacy; the regulator can not only supervise the amount of each transaction, but also supervise the sum amount for a period of time, which not only achieves the purpose of regulation, but also effectively protects user privacy; since the receiver can directly decrypt any transaction amount, we put the ciphertext legality check off-chain from the perspective of game theory, which further reduces the amount of on-chain data of the transaction.关键词：blockchain;confidential transaction;regulatable;zero-knowledge proof;homomorphic encryption319|497|2更新时间：2023-02-15
- 摘要：Aiming at solving the problem of password forgetting and leakage in the current password-based solution to the problem of blockchain private key loss, this paper proposed a Multi-Answer Protected Secret Sharing Scheme (MAPSS) based on secret problems. This scheme allows users to share a secret with multiple servers, and all servers store multiple secret questions. The scheme can not only be used to recover the private key of the blockchain, but also support the retrieval strategy against forgetting and leakage, and the scheme does not require public key infrastructure and is efficient. After that, the user only needs to provide a threshold number of answers to a threshold number of servers to reconstruct the secret. Finally, this paper not only proved that the security of the scheme is based on the Threshold Parallel One-More Diffie-Hellman (TP-OMDH) assumption under the random oracle model, but also implemented the system prototype of the scheme, which proved the practicality of the scheme.关键词：blockchain private key;secret problems;secret sharing;multi-answer protected secret sharing scheme134|990|2更新时间：2023-02-15
- 摘要：In recent years, text level fine-grained sentiment analysis has received more and more attention, and its role in medical text is also growing. Compared with coarse-grained sentiment analysis, aspect-level sentiment analysis can distinguish each specific aspect word of the medical text, and can get the emotional information expressed by each aspect word. Aspect-level sentiment analysis task need consider the interaction between aspect words and sentiment words. Medical texts can be used as both aspect words and sentiment words. So in this paper, we propose an aspect-level sentiment analysis model with Contextual Position Latent Information. At the same time, the context words related to the sentiment polarity judgment of specific aspect words in the medical text are generally located near the aspect words. Moreover, due to the difference in the number of words in the context of medical aspect words, the attribute of word embedding vector representation may change, making the relative position of aspect words different. Therefore, this paper proposes a new context location adjustment function to enhance the pertinence of the sentiment polarity words related to the designated aspect words, and reduce the interference of the number of words on both sides of the aspect words on the judgment of sentiment polarity. At the same time, a linear conditional random field model is introduced to help establish the vector representation model of aspect words in order to express the aspect words containing the emotional information of a specific aspect. Finally, the focus loss function is used to train the model parameters to deal with the class imbalance of sentiment analysis in medical texts.关键词：natural language process;aspect-level sentiment analysis;context weight adjustment;focal loss function428|624|0更新时间：2023-02-15
- 摘要：Electronic medical records of breast cancer contain rich information about breast cancer diagnosis and treatment, from which features can be extracted to construct a knowledge graph. This is helpful for the construction of both intelligent user interface and computer-assisted decision support systems of breast cancer. In this paper, a case study on feature extraction, knowledge graph construction, and application is carried out using breast cancer ultrasound examination reports. First, the entities and conceptual structure of the knowledge graph for breast cancer diagnosis are defined based on the diagnosis and treatment guidelines and clinical expert experiences. Then, based on the BiLSTM-CRF model and attention mechanism, feature extraction from ultrasonic examination reports is carried out. The results show that the accuracy rate, the recall rate, and the F1 value reach more than 96%. Finally, the knowledge graph is constructed and applied to data query, visual analysis, and computer-assisted diagnosis. The results show that it has great application value in helping improve the quality of electronic medical records, the effects of clinical diagnosis and treatment, and the performance of medical decision-making.关键词：breast cancer;knowledge graph;entity extraction;visualization;decision support486|607|1更新时间：2023-02-15
- 摘要：The task of image geo-localization refers to the prediction of geographic label for given image. Such prediction was achieved by matching the input with existing geographic labelled image in the database. Due to the lack of ground images with geographic labels, databases were often established by satellite images. However, the huge perspective difference between satellite images and ground images brought great challenges to image matching. In this paper, we proposed a new conditional generative adversarial network (CGAN) called Crossview Attention Seq (CAS) for cross-view image conversion. CAS achieved better effect through semantic segmentation of image, meanwhile we introduced spatial attention mechanism to suppress the noise in transformation. The auxiliary information was input to the framework to optimize the parameters. The image matching framework was built based on Siamese network architecture. In addition, the improved loss function was integrated into the training process. Compared with Triplet loss, it greatly improved the optimization effect. Experimental results prove the effectiveness and superiority of our method on two classical datasets. Compared with the existing methods, our model is more accurate in geo-location and more robust to low quality data.关键词：image geo-localization;cross view image transformation;Siamese network515|987|2更新时间：2023-02-15
- 摘要：RGB information and depth information are both crucial feature cues in RGB-D saliency detection. However, the current RGB-D saliency detection models are unable to handle the interaction of multi-scale features and the fusion of multi-modal features efficiently, so they are limited in understanding natural scenes. To alleviate such shortcomings, we propose a synergistic attention network for RGB-D saliency detection (SANet), and introduce a general multi-modal learning strategy, termed Guidance and Teaching Strategy. The implicit guidance strategy works in the multi-scale feature extraction stage of the encoder, while the explicit teaching strategy works in the feature fusion stage of the decoder, so as to realize the phased learning process of encoding and decoding stages. Extensive experiments are conducted on four saliency detection benchmarks, and the results verify the proposed method performs better than the 18 cutting-edge RGB-D saliency detection models in terms of four metrics.关键词：Visual perception mechanism; Synergistic attention; Saliency detection; RGB-D saliency detection203|748|0更新时间：2023-02-15
- 摘要：In this work, an attention-based 3D convolutional neural network (ACN) is proposed for continuous sign language recognition in complex background. Firstly, the sign language video containing complex background is preprocessed with the background removal module. Then, the spatio-temporal fusion information is extracted by 3D-ResNet (3D residual convolutional neural network) based on spatial attention mechanism. Finally, the long short-term memory (LSTM) network combined with the time attention mechanism is used for sequence learning to obtain the final recognition result. Extensive experiments show that the algorithm performs well on the large-scale Chinese continuous sign language dataset CSL100. The algorithm shows good generalization performance facing different complex backgrounds, and the spatio-temporal attention mechanism introduced by the model is effective.关键词：continuous sign language recognition;complex background;attention mechanism;long short-term memory（LSTM）447|807|1更新时间：2023-02-15
- 摘要：On the task of document recommendation in the architectural field, in order to better understand user preferences, a multi task learning method, KGRP (unifying knowledge graph learning and recommendation: for a better user preferences) is proposed, which embeds the knowledge graph and document recommendation tasks for joint learning. We designed a cross compression unit for KGRP, which can explicitly model the high-level interaction between document features and entity features, supplement the information of documents and entities, and enable the two tasks to share more feature information. The experimental results on the document dataset in the architectural field show that the KGRP algorithm has good recommendation performance, and can recommend appropriate documents according to the user’s interaction behavior and preference model.关键词：recommendation system;knowledge graph;joint learning;interpretability144|366|0更新时间：2023-02-15
- 摘要：Unconscious face recognition is an identification method that automatically capture and identify face images under the user’s perception-free condition, including three stages of human-machine unconscious interaction, target extraction, and face recognition. Unconscious face recognition can realize low-interaction, contactless, and highly parallelized authentication, and is suitable for unconscious scenarios with real-time requirements. The contents and principles of multiple technologies are summarized, including face registration and database construction technologies which can improve the reliability of unconscious face recognition system, unconscious face image acquisition technologies which can reduce data redundancy, and multi-scale face extraction and recognition technology based on deep learning. Then the applications and challenges of unconscious face recognition at the current stage are analyzed. With the further development of the related technologies, unconscious face recognition technology will be widely applied in security, payment, and service fields by virtue of its advantages of unconscious and parallelism.关键词：unconscious face recognition;touchless interaction;automatic image collection;identification365|870|1更新时间：2023-02-15
- Tel：027-68756952 Email：whdz@whu.edu.cn
- Technical support is provided by Beijing Founder electronics co., LTD 京ICP备09064830号-19
京公网安备11010802024621 - It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
- Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.
- Total Visits：1304640 Visits Today：327
Bulk Citation
0