检索结果-内蒙古大学图书馆

Harry: A Scalable SIMD-based Multi-literal Pattern Matching Engine for Deep Packet Inspection

学校读者我要写书评

暂无评论

Harry: A Scalable SIMD-based Multi-literal Pattern Matching ...

IEEE Annual Joint Conference: INFOCOM, IEEE computer and Communications Societies

作者： Hao Xu Harry Chang Wenjun Zhu Yang Hong Geoff Langdale Kun Qiu Jin Zhao School of Computer Science Fudan University China Shanghai Key Laboratory of Intelligent Information Processing Shanghai China Intel Asia-Pacific Research & Development Ltd. Shanghai China

Deep Packet Inspection (DPI) is a significant network security technique. It examines traffic workloads by searching for specific rules. Since every byte of packets needs to be examined by many literal rules, multi-literal matching becomes the performance bottleneck of DPI. FDR, the fastest multi-literal matching engine on CPUs, takes advantage of Single-Instruction-Multiple-Data (SIMD) to alleviate this bottleneck and achieves a performance boost over the widely-used Aho-Corasick (AC) algorithm. However, FDR does not deeply exploit the data-level parallelism of SIMD and its SIMD vector utilization is only 50%. Besides, limited by certain SIMD shift instructions, it cannot benefit from advanced SIMD instruction sets. To overcome these issues, we propose Harry, a scalable and SIMD-based multi-literal matching engine. Harry adopts a column-vector-based matching algorithm to improve the data-level parallelism and SIMD vector utilization. To support the algorithm, it takes two encoding methods to compress the mask table. Also, it utilizes shuffle instruction to implement shift. We implement Harry on commodity CPU and evaluate it with real network traffic and DPI rules. Our evaluation shows that Harry reaches a throughput of 30∼70Gbit/s, up to 52x that of AC and 2.09x of FDR. It has been successfully deployed in Hyperscan.

关键词：

AUTOMATIC CHINESE NATIONAL PENTATONIC MODES RECOGNITION USING CONVOLUTIONAL NEURAL NETWORK 23

学校读者我要写书评

暂无评论

AUTOMATIC CHINESE NATIONAL PENTATONIC MODES RECOGNITION USIN...

23rd International Society for Music information Retrieval Conference, ISMIR 2022

作者： Wang, Zhaowen Che, Mingjin Yang, Yue Meng, Wenwu Li, Qinyu Xia, Fan Li, Wei Department of Music AI and Information Technology Central Conservatory of Music China College of Experimental Art Sichuan Conservatory of Music China School of Computer Science and Technology Fudan University China Shanghai Key Laboratory of Intelligent Information Processing Fudan University China

ISBN: (纸本)9781732729926

Chinese national pentatonic modes, with five tones of Gong, Shang, Jue, Zhi and Yu as the core, play an essential role in traditional Chinese music culture. After the early twentieth century, with the development of new Chinese music, the ancient Chinese theory of scales gradually developed into a new pentatonic modes theory under the influence of western music. In this paper, we briefly introduce our self-built CNPM (Chinese National Pentatonic Modes) Dataset, then design residual convolutional neural network models to identify which TongGong system the mode belongs, the pitch of tonic, the mode pattern and the mode type from audio signals, in combination with musical domain knowledge. We use both single-task and multi-task models with three strategies for identification, and compare them with a simple template-based baseline method. In experiments, we use seven accuracy metrics to evaluate the models. The results on identifying both the tonic pitch and the pattern of mode correctly achieve an average accuracy of 69.65%. As an initial research on automatic Chinese national pentatonic modes recognition, this work will contribute to the development of multicultural music information retrieval, computational ethnomusicology and five-tone music therapy. © Z. Wang, M. Che, Y. Yang, W. Meng, Q. Li, F. Xia and W. Li.

关键词： Convolutional neural networks

ME-D2N: Multi-Expert Domain Decompositional Network for Cross-Domain Few-Shot Learning

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Fu, Yuqian Xie, Yu Fu, Yanwei Chen, Jingjing Jiang, Yu-Gang Shanghai Key Lab of Intelligent Information Processing School of Computer Science Fudan University China School of Data Science Fudan University China

Recently, Cross-Domain Few-Shot Learning (CD-FSL) which aims at addressing the Few-Shot Learning (FSL) problem across different domains has attracted rising attention. The core challenge of CD-FSL lies in the domain gap between the source and novel target datasets. Though many attempts have been made for CD-FSL without any target data during model training, the huge domain gap makes it still hard for existing CD-FSL methods to achieve very satisfactory results. Alternatively, learning CD-FSL models with few labeled target domain data which is more realistic and promising is advocated in previous work [13]. Thus, in this paper, we stick to this setting and technically contribute a novel Multi-Expert Domain Decompositional Network (ME-D2N). Concretely, to solve the data imbalance problem between the source data with sufficient examples and the auxiliary target data with limited examples, we build our model under the umbrella of multi-expert learning. Two teacher models which can be considered to be experts in their corresponding domain are first trained on the source and the auxiliary target sets, respectively. Then, the knowledge distillation technique is introduced to transfer the knowledge from two teachers to a unified student model. Taking a step further, to help our student model learn knowledge from different domain teachers simultaneously, we further present a novel domain decomposition module that learns to decompose the student model into two domain-related sub-parts. This is achieved by a novel domain-specific gate that learns to assign each filter to only one specific domain in a learnable way. Extensive experiments demonstrate the effectiveness of our method. Copyright © 2022, The Authors. All rights reserved.

关键词： Distillation

FedGAC: Graph Federated Learning with Gradients Aggregation Calibration for Non-IID and Long-Tailed Data 22

学校读者我要写书评

暂无评论

FedGAC: Graph Federated Learning with Gradients Aggregation ...

22nd IEEE International Symposium on Parallel and Distributed processing with Applications, ISPA 2024

作者： Li, Jie Wang, Jinyan Deng, Rongbin Yan, Dongqi Li, Qiyu Guangxi Normal University Key Lab of Education Blockchain and Intelligent Technology Ministry of Education Guilin541004 China Guangxi Normal University Guangxi Key Lab of Multi-Source Information Mining and Security Guilin541004 China Guangxi Normal University School of Computer Science and Engineering Guilin541004 China

ISBN: (纸本)9798331509712

Federated graph learning (FGL) combines the advantages of graph neural networks and federated learning, extracting valuable information from decentralized graph data while preserving data privacy. Existing FGL methods typically consider mitigating the effect of non-IID under the class balanced assumption. However, in many real-world scenarios, the data class distribution often exhibits a long-tailed distribution, which can lead to the model being severely biased towards the head classes, thus affecting the overall performance of the global model. To address the challenges posed by non-IID and long-tailed distribution data in FGL, this paper proposes a method called FedGAC to train a globally balanced model. In the local stage, FedGAC employs feature-structure decoupled graph attention network to train client models, introduces class-importance balanced strategy to mitigate class imbalance issues, and leverages knowledge relearning strategy to enhance model generalization. In the global stage, FedGAC employs an adaptive gradients aggregation strategy that dynamically adjusts balancing factors based on the degree of class imbalance in each client's data, thereby optimizing the aggregation process. Additionally, it utilizes global class gradients to calibrate learnable features. Experimental results demonstrate that, compared to the state-of-the-art federated learning methods, FedGAC achieves superior performance on highly non-IID and long-tailed distribution datasets. © 2024 IEEE.

关键词： Federated learning

Modeling Layout Reading Order as Ordering Relations for Visually-rich Document Understanding

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Zhang, Chong Tu, Yi Zhao, Yixi Yuan, Chenshu Chen, Huan Zhang, Yue Chai, Mingxu Guo, Ya Zhu, Huijia Zhang, Qi Gui, Tao School of Computer Science Fudan University Shanghai China Shanghai Key Laboratory of Intelligent Information Processing Shanghai China Ant Tiansuan Security Lab Ant Group Hangzhou China School of Statistics and Data Science Nankai University Tianjin China Institute of Modern Languages and Linguistics Fudan University Shanghai China

Modeling and leveraging layout reading order in visually-rich documents (VrDs) is critical in document intelligence as it captures the rich structure semantics within documents. Previous works typically formulated layout reading order as a permutation of layout elements, i.e. a sequence containing all the layout elements. However, we argue that this formulation does not adequately convey the complete reading order information in the layout, which may potentially lead to performance decline in downstream VrD tasks. To address this issue, we propose to model the layout reading order as ordering relations over the set of layout elements, which have sufficient expressive capability for the complete reading order information. To enable empirical evaluation on methods towards the improved form of reading order prediction (ROP), we establish a comprehensive benchmark dataset including the reading order annotation as relations over layout elements, together with a relation-extraction-based method that outperforms previous methods. Moreover, to highlight the practical benefits of introducing the improved form of layout reading order, we propose a reading-order-relation-enhancing pipeline to improve model performance on any arbitrary VrD task by introducing additional reading order relation inputs. Comprehensive results demonstrate that the pipeline generally benefits downstream VrD tasks: (1) with utilizing the reading order relation information, the enhanced downstream models achieve SOTA results on both two task settings of the targeted dataset;(2) with utilizing the pseudo reading order information generated by the proposed ROP model, the performance of the enhanced models has improved across all three models and eight cross-domain VrD-IE/QA task settings without targeted optimization. Copyright © 2024, The Authors. All rights reserved.

关键词： Semantics

Wave-SAN: Wavelet based Style Augmentation Network for Cross-Domain Few-Shot Learning

学校读者我要写书评

暂无评论

arXiv 2022年

Previous few-shot learning (FSL) works mostly are limited to natural images of general concepts and categories. These works assume very high visual similarity between the source and target classes. In contrast, the recently proposed crossdomain few-shot learning (CD-FSL) aims at transferring knowledge from general nature images of many labeled examples to novel domain-specific target categories of only a few labeled examples. The key challenge of CD-FSL lies in the huge data shift between source and target domains, which is typically in the form of totally different visual styles. This makes it very nontrivial to directly extend the classical FSL methods to address the CD-FSL task. To this end, this paper studies the problem of CD-FSL by spanning the style distributions of the source dataset. Particularly, wavelet transform is introduced to enable the decomposition of visual representations into low-frequency components such as shape and style and highfrequency components e.g., texture. To make our model robust to visual styles, the source images are augmented by swapping the styles of their low-frequency components with each other. We propose a novel Style Augmentation (StyleAug) module to implement this idea. Furthermore, we present a Self-Supervised Learning (SSL) module to ensure the predictions of style-augmented images are semantically similar to the unchanged ones. This avoids the potential semantic drift problem in exchanging the styles. Extensive experiments on two CD-FSL benchmarks show the effectiveness of our method. Our codes and models will be released. Copyright © 2022, The Authors. All rights reserved.

关键词： Textures

Hi-EF: Benchmarking Emotion Forecasting in Human-interaction

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Wang, Haoran Mai, Xinji Tao, Zeng Wang, Yan Yu, Jiawen Zhou, Ziheng Tong, Xuan Yan, Shaoqi Zhao, Qing Gao, Shuyong Zhang, Wenqiang Shanghai Engineering Research Center of AI & Robotics Academy for Engineering & Technology Fudan University Shanghai China Shanghai Key Lab of Intelligent Information Processing School of Computer Science Fudan University Shanghai China Engineering Research Center of Robotics Ministry of Education Academy for Engineering & Technology Fudan University Shanghai China School of Information Science and Technology Fudan University Shanghai China

Affective Forecasting, a research direction in psychology that predicts individuals' future emotions, is often constrained by numerous external factors like social influence and temporal distance. To address this, we transform Affective Forecasting into a Deep Learning problem by designing an Emotion Forecasting paradigm based on two-party interactions. We propose a novel Emotion Forecasting (EF) task grounded in the theory that an individual's emotions are easily influenced by the emotions or other information conveyed during interactions with another person. To tackle this task, we have developed a specialized dataset, Humaninteraction- based Emotion Forecasting (Hi-EF), which contains 3069 two-party Multilayered-Contextual Interaction Samples (MCIS) with abundant affectiverelevant labels and three modalities. Hi-EF not only demonstrates the feasibility of the EF task but also highlights its potential. Additionally, we propose a methodology that establishes a foundational and referential baseline model for the EF task and extensive experiments are provided. The dataset and code is available HERE. Copyright © 2024, The Authors. All rights reserved.

关键词： Prediction models

(Image Present) ForgerySleuth: Empowering Multimodal Large Language Models for Image Manipulation Detection

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Sun, Zhihao Jiang, Haoran Chen, Haoran Cao, Yixin Qiu, Xipeng Wu, Zuxuan Jiang, Yu-Gang Shanghai Key Lab of Intell. Info. Processing School of CS Fudan University China Shanghai Collaborative Innovation Center of Intelligent Visual Computing China School of Computer Science Fudan University China

(Image Present) Multimodal large language models have unlocked new possibilities for various multimodal tasks. However, their potential in image manipulation detection remains unexplored. When directly applied to the IMD task, M-LLMs often produce reasoning texts that suffer from hallucinations and overthinking. To address this, in this work, we propose ForgerySleuth, which leverages M-LLMs to perform comprehensive clue fusion and generate segmentation outputs indicating specific regions that are tampered with. Moreover, we construct the ForgeryAnalysis dataset through the Chain-of-Clues prompt, which includes analysis and reasoning text to upgrade the image manipulation detection task. A data engine is also introduced to build a larger-scale dataset for the pre-training phase. Our extensive experiments demonstrate the effectiveness of ForgeryAnalysis and show that ForgerySleuth significantly outperforms existing methods in generalization, robustness, and explainability. The resource is available at https://***/sunzhihao18/ForgerySleuth. Copyright © 2024, The Authors. All rights reserved.

关键词： Large datasets

Towards End-to-End Unsupervised Saliency Detection with Self-Supervised Top-Down Context

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Song, Yicheng Gao, Shuyong Xing, Haozhe Cheng, Yiting Wang, Yan Zhang, Wenqiang Shanghai Key Lab of Intelligent Information Processing School of Computer Science Fudan University Shanghai China Keenon Robotics Co. Ltd. Shanghai China Academy for Engineering & Technology Fudan University Shanghai China

Unsupervised salient object detection aims to detect salient objects without using supervision signals eliminating the tedious task of manually labeling salient objects. To improve training efficiency, end-to-end methods for USOD have been proposed as a promising alternative. However, current solutions rely heavily on noisy handcraft labels and fail to mine rich semantic information from deep features. In this paper, we propose a self-supervised end-to-end salient object detection framework via top-down context. Specifically, motivated by contrastive learning, we exploit the self-localization from the deepest feature to construct the location maps which are then leveraged to learn the most instructive segmentation guidance. Further considering the lack of detailed information in deepest features, we exploit the detail-boosting refiner module to enrich the location labels with details. Moreover, we observe that due to lack of supervision, current unsupervised saliency models tend to detect non-salient objects that are salient in some other samples of corresponding scenarios. To address this widespread issue, we design a novel Unsupervised Non-Salient Suppression (UNSS) method developing the ability to ignore non-salient objects. Extensive experiments on benchmark datasets demonstrate that our method achieves leading performance among the recent end-to-end methods and most of the multi-stage solutions. The code is available. Copyright © 2023, The Authors. All rights reserved.

关键词： Object detection