检索结果-内蒙古大学图书馆

BWG: An IOC Identification Method for Imbalanced Threat Intelligence Datasets

学校读者我要写书评

暂无评论

BWG: An IOC Identification Method for Imbalanced Threat Inte...

IEEE International Conference on Trust, Security and Privacy in Computing and Communications (TrustCom)

作者： Juncheng Lu Yiyang Zhao Yan Wang Jiyuan Cui Sanfeng Zhang School of Cyber Science and Engineering Southeast University Nanjing China Key Laboratory of Computer Network and Information Integration Southeast University Ministry of Education Nanjing China

ISBN: (数字)9798331506209

ISBN: (纸本)9798331506216

APT attacks are becoming increasingly complex and stealthy. To effectively counter APT attacks, modelling threat intelligence data based on graphs, identifying Indicators of Compromise (IOC) nodes, and providing early warnings have become new research hotspots. However, the problem of node category imbalance in such graph datasets restricts the identification capabilities of these methods. Therefore, this paper proposes a supervised graph data augmentation method. In the training phase, graph disentangled representation learning is utilized to perform feature embedding for minority class nodes, effectively alleviating the sparsity problem faced by traditional methods and effectively integrating neighbourhood information of minority class nodes at a higher semantic level. Additionally, two loss functions designed based on link prediction and prototype constraints enhance node type consistency and semantic consistency, respectively. Experimental results on the APT and PDNS datasets demonstrate that the proposed method outperforms other baseline models in identification performance; even in highly imbalanced scenarios, it surpasses the second-best model.

关键词： Training Data privacy Computational modeling Semantics Disentangled representation learning Prototypes Data augmentation Data models Security

A Novel Classical-Quantum Transfer Learning Framework for Image Recognition

学校读者我要写书评

暂无评论

SSRN

SSRN 2024年

作者： Ruan, Banyao Liu, Zhihao Li, Xi School of Computer Science and Engineering Southeast University Nanjing211189 China Key Laboratory of Computer Network and Information Integration Southeast University Ministry of Education Nanjing211189 China

With the advent of the era of Noisy Intermediate-Scale quantum (NISQ), quantum processors have demonstrated clear quantum supremacy, marking a new milestone in the development of quantum computing. At the same time, classical deep learning algorithms have been widely used in many fields such as image recognition and natural language processing. In this context, the Classical-Quantum transfer learning (CQ-TL) method, which combines the advantages of quantum computing and classical deep learning algorithms, emerged. In view of the problem of unclear quantum advantage in the existing CQ-TL framework, this paper proposes a novel CQ-TL framework based on amplitude encoding and effective learning algorithm is given. It not only has the flexibility of the existing framework, but also reduces the number of classical parameters, and can better demonstrate quantum advantages. Based on the CQ-TL framework, we designed and implemented two CQ-TL neural network models: named Transfer Learning Quantum Neural network (TLQNN) and Transfer Learning Quantum Convolutional Neural network (TLQCNN). TLQNN is structurally similar to a multi-layer feedforward neural network, and it achieves feature extraction through repeated quantum unit layers (ansatz). TLQCNN, on the other hand, employs quantum convolution and quantum pooling operations, which can effectively alleviate the effects of the barren plateau phenomenon, making it highly trainable and possessing great potential for practical applications. In cross-experiments with three benchmark datasets (MNIST, Fashion-MNIST and CIFAR10) and three source models (ResNet18, ResNet50 and DenseNet121), the two models have exceeded the benchmark classical classifier in many performance measures such as accuracy, convergence, stability and generalization capability. Our research shows that the proposed novel CQ-TL framework and its models have potential quantum advantages in the NISQ era, providing strong support for the practical application of qu

关键词： Learning algorithms

Benchmarking Temporal Reasoning and Alignment Across Chinese Dynasties

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Wang, Zhenglin Wu, Jialong Li, Pengfei Jiang, Yong Zhou, Deyu School of Computer Science and Engineering Key Laboratory of Computer Network and Information Integration Ministry of Education Southeast University China Tongyi Lab Alibaba Group China

Temporal reasoning is fundamental to human cognition and is crucial for various real-world applications. While recent advances in Large Language Models have demonstrated promising capabilities in temporal reasoning, existing benchmarks primarily rely on rule-based construction, lack contextual depth, and involve a limited range of temporal entities. To address these limitations, we introduce Chinese Time Reasoning (CTM), a benchmark designed to evaluate LLMs on temporal reasoning within the extensive scope of Chinese dynastic chronology. CTM emphasizes cross-entity relationships, pairwise temporal alignment, and contextualized and culturally-grounded reasoning, providing a comprehensive evaluation. Extensive experimental results reveal the challenges posed by CTM and highlight potential avenues for improvement. Copyright © 2025, The Authors. All rights reserved.

关键词： Benchmarking

Calibration bottleneck: over-compressed representations are less calibratable 24

学校读者我要写书评

暂无评论

Calibration bottleneck: over-compressed representations are ...

Proceedings of the 41st International Conference on Machine Learning

作者： Deng-Bao Wang Min-Ling Zhang School of Computer Science and Engineering Southeast University Nanjing China and Key Lab. of Computer Network and Information Integration (Southeast University) MOE China

Although deep neural networks have achieved remarkable success, they often exhibit a significant deficiency in reliable uncertainty calibration. This paper focus on model calibratability, which assesses how amenable a model is to be well recalibrated post-hoc. We find that the widely used weight decay regularizer detrimentally affects model calibratability, subsequently leading to a decline in final calibration performance after post-hoc calibration. To identify the underlying causes leading to poor calibratability, we delve into the calibratability of intermediate features across the hidden layers. We observe a U-shaped trend in the calibratability of intermediate features from the bottom to the top layers, which indicates that over-compression of the top representation layers significantly hinders model calibratability. Based on the observations, this paper introduces a weak classifier hypothesis, i.e., given a weak classification head that has not been over-trained, the representation module can be better learned to produce more calibratable features. Consequently, we propose a progressively layer-peeled training (PLP) method to exploit this hypothesis, thereby enhancing model calibratability. Our comparative experiments show the effectiveness of our method, which improves model calibration and also yields competitive predictive performance.

关键词：

Improved NL2SQL based on Multi-layer Expert network

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Hao, Chenduo Zhang, Xu School of Computer Science and Engineering Southeast University China Key Laboratory of Computer Network and Information Integration Ministry of Education Southeast University China

The Natural Language to SQL (NL2SQL) technique is used to convert natural language queries into executable SQL statements. Typically, slot-filling is employed as a classification method for multi-task cases to achieve this goal. However, slot-filling can result in inaccurate SQL statement generation due to negative migration issues arising from different classification tasks. To overcome this limitation, this study introduces a new approach called Multi-Layer Expert Generate SQL (MLEG-SQL), which utilizes a dedicated multi-task hierarchical network. The lower layer of the network extracts semantic features of natural language statements, while the upper layer builds a specialized expert system for handling specific classification tasks. This hierarchical approach mitigates performance degradation resulting from different task conflicts. The proposed method was evaluated on the WiKSQL dataset and was found to be effective in generating accurate SQL statements. © 2023, CC BY.

关键词： Filling

RankMatch: A Novel Approach to Semi-Supervised Label Distribution Learning Leveraging Inter-label Correlations

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Kou, Zhiqiang Xie, Yucheng Wang, Jing Jia, Yuheng Shi, Boyu Geng, Xin MOE Key Laboratory of Computer Network and Information Integration School of Computer Science and Engineering Southeast University Nanjing China

This paper introduces RankMatch, an innovative approach for Semi-Supervised Label Distribution Learning (SSLDL). Addressing the challenge of limited labeled data, RankMatch effectively utilizes a small number of labeled examples in conjunction with a larger quantity of unlabeled data, reducing the need for extensive manual labeling in Deep Neural network (DNN) applications. Specifically, RankMatch introduces an ensemble learning-inspired averaging strategy that creates a pseudo-label distribution from multiple weakly augmented images. This not only stabilizes predictions but also enhances the model’s robustness. Beyond this, RankMatch integrates a pairwise relevance ranking (PRR) loss, capturing the complex inter-label correlations and ensuring that the predicted label distributions align with the ground truth. We establish a theoretical generalization bound for RankMatch, and through extensive experiments, demonstrate its superiority in performance against existing SSLDL methods. Copyright © 2023, The Authors. All rights reserved.

关键词： Deep neural networks

Consistent Multi-Granular Rationale Extraction for Explainable Multi-hop Fact Verification

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Si, Jiasheng Zhu, Yingjie Zhou, Deyu School of Computer Science and Engineering Key Laboratory of Computer Network and Information Integration Ministry of Education Southeast University China

The success of deep learning models on multi-hop fact verification has prompted researchers to understand the behavior behind their veracity. One possible way is erasure search: obtaining the rationale by entirely removing a subset of input without compromising the veracity prediction. Although extensively explored, existing approaches fall within the scope of the single-granular (tokens or sentences) explanation, which inevitably leads to explanation redundancy and inconsistency. To address such issues, this paper explores the viability of multi-granular rationale extraction with consistency and faithfulness for explainable multi-hop fact verification. In particular, given a pretrained veracity prediction model, both the token-level explainer and sentence-level explainer are trained simultaneously to obtain multi-granular rationales via differentiable masking. Meanwhile, three diagnostic properties (fidelity, consistency, salience) are introduced and applied to the training process, to ensure that the extracted rationales satisfy faithfulness and consistency. Experimental results on three multi-hop fact verification datasets show that the proposed approach outperforms some state-of-the-art baselines. Copyright © 2023, The Authors. All rights reserved.

关键词： Extraction

NL2SQL Enhancement through Semantic Representation Aggregation 5

学校读者我要写书评

暂无评论

NL2SQL Enhancement through Semantic Representation Aggregati...

5th International Conference on computer information Science and Application Technology, CISAT 2022

作者： Long, Lingli Zhu, Yongjin Shao, Jun Kong, Zheng Li, Jian Liu, Zejie Zhang, Xu Jiangsu Frontier Electric Technology Co. Ltd. China School of Computer Science and Engineering Key Laboratory of Computer Network and Information Integration Ministry of Education Southeast University China

ISBN: (纸本)9781510660076

Natural Language to SQL (NL2SQL) is a technology that converts natural language into executable SQL statements. For single table tasks, the dominant approach is a slot-filling-based model which actually performs multi-objective prediction. However, in existing multi-objective models, it is difficult to optimize multiple objectives simultaneously. Low correlation or conflict between various objectives can result in the model not being able to learn the parameters effectively. To effectively address this problem, this paper considers redesigning the sharing mechanism between multiple tasks and the specific network structure of individual tasks through a semantic aggregation approach in order to alleviate the above phenomenon. © 2022 SPIE.

关键词： Semantics

NL2SQL Generation with Noise Labels based on Multi-task Learning 5

学校读者我要写书评

暂无评论

NL2SQL Generation with Noise Labels based on Multi-task Lear...

2022 5th International Symposium on Big Data and Applied Statistics, ISBDAS 2022

作者： Long, Lingli Zhu, Yongjin Shao, Jun Kong, Zheng Li, Jian Xiang, Yanzheng Zhang, Xu Jiangsu Frontier Electric Technology Co. Ltd. China School of Computer Science and Engineering Key Laboratory of Computer Network and Information Integration Ministry of Education Southeast University China

With the rapid development of artificial intelligence technology, semantic recognition technology is becoming more and more mature, providing the preconditions for the development of natural language to SQL (NL2SQL) technology. In the latest research on NL2SQL, the use of pre-trained models as feature extractors for natural language and table schema has led to a very significant improvement in the effectiveness of the models. However, the current models do not take into account the degradation of the noisy labels on the overall SQL statement generation. It is crucial to reduce the impact of noisy labels on the overall SQL generation task and to maximize the return of accurate answers. To address this issue, we propose a restrictive constraint-based approach to mitigate the impact of noise-labeled labels on other tasks. In addition, parameter sharing approach is used in noiseless-labeled labels to capture each part's correlations and improve the robustness of the model. In addition, we propose to use Kullback-Leibler divergence to constrain the discrepancy between hard and soft constrained coding of noisy labels. Our model is compared with some recent state-of-the-art methods, and experimental results show a significant improvement over the approach in this paper. © Published under licence by IOP Publishing Ltd.

关键词： Semantics