检索结果-内蒙古大学图书馆

SLMP: A Scientific Literature Management Platform Based on Large Language Models

学校读者我要写书评

暂无评论

SLMP: A Scientific Literature Management Platform Based on L...

IEEE International Conference on big Knowledge (ICBK)

作者： Menghao Guo Jinling Jiang Fan Wu Shanxin Sun Chen Zhang Wenhui Li Zeyi Sun Guangyong Chen Xindong Wu Research Center for Life Sciences Computing Zhejiang Lab Hangzhou China Research Center for Data Hub and Security Zhejiang Lab Hangzhou China Research Center for High Efficiency Computing System Zhejiang Lab Hangzhou China Key Laboratory of Knowledge Engineering With Big Data Hefei University of Technology Hefei China

ISBN: (数字)9798331508821

ISBN: (纸本)9798331508838

This paper presents a Scientific Literature Management Platform (SLMP, demo link 1 ) based on large language models (LLMs). The platform consists of four modules: literature management, literature extraction, literature retrieval, and question answering. The core techniques used to support the four modules across the platform include a fine-tuned model PaperExtractGPT and a continual pre-training model ChatPaperGPT based on ChatGLM 2 using the data from scientific research literature, responsible for information extraction and communication, respectively. Due to their powerful capabilities in natural language understanding and generation, LLMs can understand complex scientific concepts based on the provided contexts, and thus generate high-quality texts and conduct in-depth information retrieval and question answering. Our platform can help researchers manage and utilize literature more effectively and efficiently for finding relevant literature, obtaining required information, and generating new knowledge.

关键词： Training Large language models Knowledge graphs Information retrieval data models Question answering (information retrieval)

Graph-Augmented Contrastive Clustering for Time Series

学校读者我要写书评

暂无评论

SSRN 2023年

作者： Zhang, Qin Liang, Zhuoluo Ngueilbaye, Alladoumbaye Zhang, Peng Chen, Junyang Chen, Xiaojun Huang, Joshua Zhexue Big Data Institute College of Computer Science and Software Engineering Shenzhen University Shenzhen518060 China National Engineering Laboratory for Big Data System Computing Technology Shenzhen University Shenzhen518060 China Cyberspace Institute of Advanced Technology Guangzhou University Guangzhou510006 China

The recent emergence of time series contrastive clustering methods can be broadly categorized into two classes. The first class uses contrastive learning to learn universal representations for time series. Though they perform well in various downstream tasks, such methods disregard the important categorical information and clustering objective, leading to unsuitable representations of clustering tasks. The second class incorporates contrastive learning into the clustering objective. The potential connections and structures between the time series data are not fully explored during contrastive learning. To this end, we propose a graph-augmented contrastive learning framework called the "Time Series Graph-augmented Contrastive Clustering (TSGCC) method." We observed that the original samples should be similar to their augmentations and other samples in the same cluster. Hence, we used a weighted $KNN$ graph to build positive and negative sample pairs for contrastive learning. Subsequently, we projected the instance samples into a feature space with the dimensionality number of clusters and learned cluster-friendly features and cluster assignments by iteratively optimizing the contrastive loss. Experimental results demonstrate that TSGCC outperforms 16 advanced time-series clustering methods on 36 challenging UCR benchmarks, achieving the best clustering results on 12 datasets and the highest average rank (2.83) of RI overall methods. © 2023, The Authors. All rights reserved.

关键词： Time series

Robin: A Novel Method to Produce Robust Interpreters for Deep Learning-Based Code Classifiers

学校读者我要写书评

暂无评论

Robin: A Novel Method to Produce Robust Interpreters for Dee...

IEEE International Conference on Automated Software engineering (ASE)

作者： Zhen Li Ruqian Zhang Deqing Zou Ning Wang Yating Li Shouhuai Xu Chen Chen Hai Jin School of Cyber Science and Engineering Huazhong University of Science and Technology Wuhan China Services Computing Technology and System Lab Hubei Key Laboratory of Distributed System Security Cluster and Grid Computing Lab National Engineering Research Center for Big Data Technology and System Hubei Engineering Research Center on Big Data Security Department of Computer Science University of Colorado Colorado Springs USA Center for Research in Computer Vision University of Central Florida USA School of Computer Science and Technology Huazhong University of Science and Technology Wuhan China

Deep learning has been widely used in source code classification tasks, such as code classification according to their functionalities, code authorship attribution, and vulnerability detection. Unfortunately, the black-box nature of deep learning makes it hard to interpret and understand why a classifier (i.e., classification model) makes a particular prediction on a given example. This lack of interpretability (or explainability) might have hindered their adoption by practitioners because it is not clear when they should or should not trust a classifier's prediction. The lack of interpretability has motivated a number of studies in recent years. However, existing methods are neither robust nor able to cope with out-of-distribution examples. In this paper, we propose a novel method to produce Robust interpreters for a given deep learning-based code classifier; the method is dubbed Robin. The key idea behind Robin is a novel hybrid structure combining an interpreter and two approximators, while leveraging the ideas of adversarial training and data augmentation. Experimental results show that on average the interpreter produced by Robin achieves a 6.11% higher fidelity (evaluated on the classifier), 67.22% higher fidelity (evaluated on the approximator), and 15.87x higher robustness than that of the three existing interpreters we evaluated. Moreover, the interpreter is 47.31% less affected by out-of-distribution examples than that of LEMNA.

关键词：

CTGDiff: A Conditional Diffusion Model for Cardiotocography Signal Synthesis

学校读者我要写书评

暂无评论

CTGDiff: A Conditional Diffusion Model for Cardiotocography ...

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Xiaoqing Li Pufan Cai Yu Lu Shijie Shi Liangkun Ma Xianghua Fu College of Big Data and Internet Shenzhen Technology University Shenzhen China National Engineering Laboratory for Big Data System Computing Technology Shenzhen University Shenzhen China Department of Surgery Yong Loo Lin School of Medicine National University of Singapore Singapore School of Applied Technology Shenzhen University Shenzhen China Department of Obstetrics and Gynecology Peking Union Medical College Hospital Beijing China

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

The analysis of Cardiotocography (CTG) signals is often hindered by challenges such as limited data availability and label imbalance, which can undermine the performance of deep learning models. To address these issues, we present CTGDiff, a novel conditional diffusion model designed for generating synthetic Fetal Heart Rate (FHR) and Uterine Contraction (UC) signals. CTGDiff leverages both Phase-Rectified Signal Averaging (PRSA) spectrograms and UC as conditioning inputs for FHR, and integrates time encoding, condition generation from PRSA features, and residual blocks with dilated convolutions to capture both temporal dynamics and long-range dependencies. Extensive experiments, both qualitative and quantitative, demonstrate the model’s ability to synthesize high-quality CTG signals. In comparison with GANs and image-based diffusion models, CTGDiff achieves superior signal fidelity and distribution similarity for FHR, as indicated by metrics such as a 0.004 maximum mean deviation (MMD), 0.646 percent root mean square difference (PRD), 3.951 relative entropy (RE), and 0.291 Frechet distance (FD). Expert evaluations confirm that the model can generate both normal and abnormal CTG signals with high accuracy, conditioned on specific input data. These results underscore the potential of diffusion models for a wide range of applications in biomedical time series analysis, including signal synthesis, imputation, and noise reduction.

关键词： Measurement Fetal heart rate Biological system modeling Time series analysis Diffusion models data models Signal synthesis Cardiography Speech processing Spectrogram

Personalized Federated Learning with Multiple Classifier Aggregation

学校读者我要写书评

暂无评论

SSRN 2024年

作者： Zheng, Shaifeng Zhu, Qingling Lin, Qiuzhen Liu, Songbai Wong, Ka-Chun Li, Jianqiang College of Computer Science and Software Engineering Shenzhen University Shenzhen China National Engineering Laboratory for Big Data System Computing Technology Shenzhen University China Department of Computer Science City University of Hong Kong Hong Kong

Personalized federated learning (PFL) has garnered attention due to its capability to address statistical heterogeneity among clients. Typically, prevailing PFL methods aggregate a single global model for personalization, which may be inadequate for clients with diverse data distributions. Furthermore, in the local update, each private dataset is used to optimize the model independently, which increases the risk of overfitting the current data distribution and losing previously acquired knowledge, resulting in knowledge forgetting. In this study, personalized federated learning with multiple classifier aggregation is proposed. The method involves splitting the client model into its head and base, then optimizing them respectively using an alternating optimization strategy that first optimizes the head and then the base. Initially, to address the suboptimal model problem, the proposed approach aggregates multiple classifiers using data distribution and employs knowledge distillation to impart positive and negative classifier knowledge for learning the most suitable personalized model head. Additionally, to mitigate knowledge forgetting, a learnable personalization layer is introduced, and hidden loss is utilized to learn the knowledge of the global base and prevent overfitting of the model base. Empirical results demonstrate that the proposed method achieves competitive performance across various benchmarks, outperforming the majority of state-of-the-art PFL algorithms. The source code is publicly available at \href{https://***/xiaye-maker/FedMCA}{https://***/xiaye-maker/FedMCA} © 2024, The Authors. All rights reserved.

关键词： Aggregates

Dfier: A Directed Vulnerability Verifier for Ethereum Smart Contracts

学校读者我要写书评

暂无评论

SSRN 2023年

作者： Wang, Zeli Dai, Weiqi Li, Ming Choo, Kim-Kwang Raymond Zou, Deqing Chongqing Key Laboratory of Computational Intelligence Key Laboratory of Big Data Intelligent Computing Key Laboratory of Cyberspace Big Data Intelligent Security Ministry of Education College of Computer Science and Technology Chongqing University of Posts and Telecommunications Chongqing40065 China National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Hubei Engineering Research Center on Big Data Security Hubei Key Laboratory of Distributed System Security Hubei Engineering Research Center on Big Data Security School of Cyber Science and Engineering Huazhong University of Science and Technology Wuhan430074 China Department of Information Systems and Cyber Security University of Texas at San Antonio San Antonio United States

Existing smart contract vulnerability identification approaches mainly focus on complete program detection. Consequently, lots of known potentially vulnerable locations need manual verification, which is energy-exhaustive and error-prone. Fuzzing is an effective way to detect vulnerabilities. However, current fuzzers cannot effectively probe these suspicious targets because they overlook two key factors: exploring all paths toward the targets and providing high-quality directed seed inputs. This paper presents a Directed vulnerability veriFier (DFier), elaborating effective transaction sequences with directed inputs for the fuzzer, to focus on analyzing target paths and automatically detect whether the given locations are vulnerable. Specifically, DFier fuses the effectiveness of static analysis and the efficiency of fuzzing. The static analysis helps locate target paths and generate directed inputs for the fuzzer under three heuristic strategies. The first prefers inputs with the larger possibility of triggering potential vulnerabilities based on target features. The second guides the value mutation to satisfy the path constraints. The third adequately leverages execution information to help find exploits faster. Extensive experiments demonstrate that DFier is effective in verifying contract security, compared with three existing contract fuzzers (i.e., contractFuzzer, sFuzz, and conFuzzius). And the performance losses are in an acceptable range. © 2023, The Authors. All rights reserved.

关键词： Smart contract

A Novel Multi-layered Minutiae Extractor Based on OCT Fingerprints 16th

学校读者我要写书评

暂无评论

A Novel Multi-layered Minutiae Extractor Based on OCT Finge...

16th Chinese Conference on Biometric Recognition, CCBR 2022

作者： Zeng, Wenfeng Zhang, Wentian Liu, Feng Tan, Xu Li, Qin College of Computer Science and Software Engineering Shenzhen University Shenzhen China SZU Branch Shenzhen Institute of Artificial Intelligence and Robotics for Society Senzhen China The National Engineering Laboratory for Big Data System Computing Technology Shenzhen University Shenzhen China School of Software Engineering Shenzhen Institute of Information Technology Shenzhen China

ISBN: (纸本)9783031202322

Fingerprints of Optical Coherence Tomography (OCT) imaging provide 3D volume data which have the nature property of multi-layered tissue structure. This paper, for the first time, attempts to extract minutiae for OCT-based fingerprint by making full use of the merits multi-layered structure of OCT imaging and powerful convolution al neural network (CNN). In particular, a novel multi-layered feature fusion minutiae extraction network is proposed, involving a multi-layered feature extractor to integrate the rich information of multiple subsurface fingerprints, and a two-stage object detection framework using reweighted concatenation feature to detect minutiae points. Compared with the results achieved by typical traditional and learning-based methods, our method outperforms the best one by about 2% of F1-score in minutiae extraction, and achieves the improvement of 0.09% than the best one in matching performance, which further demonstrate the effectiveness and robustness of our proposed method. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Optical tomography

A Novel Adaptive Perturbation-Based Method for Long-Tailed Open Sets Recognition

学校读者我要写书评

暂无评论

SSRN 2023年

作者： Ngueilbaye, Alladoumbaye Zhang, Qin Yang, Hongye Zhang, Xiaolin Zhang, Peng Chen, Xiaojun Huang, Joshua Zhexue Big Data Institute College of Computer Science and Software Engineering Shenzhen University Shenzhen518060 China National Engineering Laboratory for Big Data System Computing Technology Shenzhen University Shenzhen518060 China College of Computer Science and Engineering Shandong University of Science and Technology Qingdao266590 China

Real-world data often have a long-tailed and open-ended distribution. A reliable practical machine learning system need to learn from the majority classes and also generalize to minority *** achieve this, acknowledge the minority and the novel class samples, and avoid misclassifying them into one of the majority classes are of great significance. Long-tailed Open Set Recognition aims to learn from the naturally distributed long-tailed and open-ended data, by using the easy-gained head class samples to learn the fundamental information, and also try to improve the out-of-distribution generalization ability by identifying the tail class and open-set class samples as novelty to avoid the misclassification. The key challenges for long-tailed open set recognition is the tail and open-set sensitivity, since they can be easily biased towards dominant classes and perform ineffectively on close-set dataset classes. Inspired by techniques on open set recognition, in this paper, we propose a Class-aware Input Preprocessing (CLAIP) approach to improve the detection of the Out-of-distribution samples by perturbing the input images. In this approach, we employ two techniques to leverage its performance. First, we introduce the middle class samples as auxiliary data during the network training process to force each category's boundaries. Second, we add a Class-aware Input Preprocessing module that can be easily plugged into the testing process by perturbing the salient regions of the input images to further improve the out-of-distribution generalization. The experimental results show that CLAIP achieves high quality and significant performance compared to the existing solutions. © 2023, The Authors. All rights reserved.

关键词： Image enhancement

Circular Decomposition and Cross-Modal Recombination for Multimodal Sentiment Analysis

学校读者我要写书评

暂无评论

Circular Decomposition and Cross-Modal Recombination for Mul...

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Haijian Liang Weicheng Xie Xilin He Siyang Song Linlin Shen School of Computer Science & Software Engineering Computer Vision Institue Shenzhen University Guangdong Key Laboratory of Intelligent Information Processing Shenzhen University School of Computing and Mathematical Sciences University of Leicester National Engineering Laboratory for Big Data System Computing Technology Shenzhen University

Multimodal Sentiment Analysis is a burgeoning research area, leveraging various modalities to predict the sentiment score. Nevertheless, previous studies have disregarded the impact of noise interference on specific modal sentiments during video recording, thereby compromising the accuracy of sentiment prediction. In this paper, we propose the Guided Circular Decomposition and Cross-Modal Recombination (GCD-CMR) model, which aims to eliminate contaminated sentiment features in a fine-grained way. To achieve this, we utilize tailored global information specific to each modality to guide the circular decomposing process in the GCD module, to produce a set of sentiment prototypes. Subsequently, in the CMR module, we align cross-modal sentiment prototypes and remove the contaminated prototypes for recombination. Experimental results on two publicly available datasets demonstrate that our model surpasses state-of-the-art models, confirming the effectiveness of our proposed method. We release the code at: https://***/nianhua20/GCD-CMR.

关键词：