检索结果-内蒙古大学图书馆

6th International Conference on Frontier Technologies of Information and Computer, ICFTIC 2024

作者： Liu, Wenyi Hu, Feng Li, Guilan Xu, Bangjian Niu, Xin College of Computer Science and Electronic Hunan University Changsha China College of Computer National University of Defense Technology Science and Technology on Parallel and Distributed Laboratory Changsha China

ISBN: (纸本)9798331541750

The multiplier is an important component of the processor's computing unit. Multiplication, multiplication, addition, and multiplication and subtraction operations are widely used in various signal processing algorithms. Based on this, this article extends the multiplication instructions of RISC-V, designs a 64-bit multiplier that combines the booth2 algorithm and wallace compression technology, and designs a deep pipeline mechanism for the multiplier to improve performance. Finally, Through logical simulation and emulation, the result output is correct. Comprehensive results show that under the 28nm cmos process, the MAC unit can reach 2.22Ghz. © 2024 IEEE.

关键词： Pipeline processing systems

来源：评论

学校读者我要写书评

暂无评论

Optimizing Depthwise Convolutions on ARMv8 Architecture 23rd

Optimizing Depthwise Convolutions on ARMv8 Architecture

引用

23rd International Conference on parallel and distributed Computing, Applications, and Technologies, PDCAT 2022

作者： Hao, Ruochen Wang, Qinglin Yin, Shangfei Zhou, Tianyang Zhang, Qingyang Mei, Songzhu Shen, Siqi Liu, Jie Science and Technology on Parallel and Distributed Processing Laboratory National University of Defense Technology Changsha410073 China College of Computer National University of Defense Technology Changsha410073 China Xiamen University Xiamen China

ISBN: (纸本)9783031299261

Depthwise convolutions are widely used in lightweight convolutional neural networks (CNNs). The performance of depthwise convolutions is mainly bounded by the memory access rather than the arithmetic operations for classic convolutions so that direct algorithms are often more efficient than indirect ones (matrix multiplication-, Winograd-, and FFT-based convolutions) with additional memory accesses. However, the existing direct implementations of depthwise convolutions on ARMv8 architectures feature a bad trade-off between register-level reuse of different tensors, which usually leads to sub-optimal performance. In this paper, we propose a new direct implementation of depthwise convolutions by means of implicit padding, register tiling, etc. Compared to the existing ones, our new implementations can incur much less communication overhead between registers and cache. Experimental results on two ARMv8 CPUs show that our implementation can averagely deliver 4.88 × performance improvement over the existing direct ones in open-source libraries. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Convolution

来源：评论

学校读者我要写书评

暂无评论

CD-Sched: An Automated Scheduling Framework for Accelerating Neural Network Training on Shared Memory CPU-DSP Platforms 23

CD-Sched: An Automated Scheduling Framework for Accelerating...

引用

Proceedings of the 2023 International Conference on Power, Communication, Computing and Networking Technologies

作者： Yuanyuan Xiao Zhiquan Lai Dongsheng Li National Key Laboratory for Parallel and Distributed Processing National University of Defense Technology China National Key Laboratory of Parallel and Distributed Processing National University of Defense Technology China

ISBN: (纸本)9781450399951

DSP holds significant potential for important applications in Deep Neural Networks. However, there is currently a lack of research focused on shared-memory CPU-DSP heterogeneous chips. This paper proposes CD-Sched, an automated scheduling framework that aims to address this gap. By predicting the latency of operators on both CPU and DSP, CD-Sched automatically schedules the computation of operators to the appropriate computing device. This scheduling optimization accelerates the computation of individual operators and ultimately improves the overall training time of neural networks. In end-to-end training tasks, CD-Sched can significantly reduce the overall training time, with an average reduction of approximately 10.77%.

关键词： computation offloading

来源：评论

学校读者我要写书评

暂无评论

High-performance Network Traffic Classification Based on Graph Neural Network 6

High-performance Network Traffic Classification Based on Gra...

引用

6th IEEE Information technology, Networking, Electronic and Automation Control Conference, ITNEC 2023

作者： Pang, Bo Fu, Yongquan Ren, Siyuan Jia, Yan Harbin Institute of Technology College of Computer Science and Technology Shenzhen China National University of Defense Technology National Key Laboratory for Parallel and Distributed Processing College of Computer Changsha China Peng Cheng Laboratory Shen Zhen China

ISBN: (纸本)9781665460033

Network traffic classification is crucial for network security and network management and is one of the most important network tasks. Current state-of-the-art traffic classifiers are based on deep learning models to automatically extract features from packet streams. Unfortunately, current approaches fail to effectively combine the structural information of traffic packets with the content features of the packets, resulting in limited classification accuracy. In this paper, we propose a graph neural network model for network traffic classification, which can well perceive the interaction feature of packets in traffic. Firstly, we design a graph structure for packets' flows to hold the interaction information between packets, which embeds both packet contents and sequence relationships into a unified graph. Secondly, we propose a graph neural network framework for graph classification to automatically learn the structural features of the packets' flows together with the packets' features. Extensive evaluation results on real-world traffic data show that the proposed model improves the prediction accuracy of improves the prediction accuracy by 2% to 37% for malicious traffic classification. © 2023 IEEE.

关键词： Graph neural networks

来源：评论

学校读者我要写书评

暂无评论

A New Pipeline for Knowledge Graph Reasoning Enhanced by Large Language Models Without Fine-Tuning

A New Pipeline for Knowledge Graph Reasoning Enhanced by Lar...

引用

2024 Conference on Empirical Methods in Natural Language processing, EMNLP 2024

作者： Chen, Zhongwu Bai, Long Li, Zixuan Huang, Zhen Jin, Xiaolong Dou, Yong National Key Laboratory of Parallel and Distributed Computing National University of Defense Technology China CAS Key Laboratory of Network Data Science and Technology Institute of Computing Technology Chinese Academy of Sciences China

ISBN: (纸本)9798891761643

Conventional Knowledge Graph Reasoning (KGR) models learn the embeddings of KG components over the structure of KGs, but their performances are limited when the KGs are severely incomplete. Recent LLM-enhanced KGR models input KG structural information into LLMs. However, they require fine-tuning on open-source LLMs and are not applicable to closed-source LLMs. Therefore, in this paper, to leverage the knowledge in LLMs without fine-tuning to assist and enhance conventional KGR models, we propose a new three-stage pipeline, including knowledge alignment, KG reasoning and entity reranking. Specifically, in the alignment stage, we propose three strategies to align the knowledge in LLMs to the KG schema by explicitly associating unconnected nodes with semantic relations. Based on the enriched KGs, we train structure-aware KGR models to integrate aligned knowledge to original knowledge existing in KGs. In the reranking stage, after obtaining the results of KGR models, we rerank the top-scored entities with LLMs to recall correct answers further. Experiments show our pipeline can enhance the KGR performance in both incomplete and general situations. © 2024 Association for Computational Linguistics.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Transformer in reinforcement learning for decision-making:a survey

引用

Frontiers of Information technology & Electronic Engineering 2024年第6期25卷 763-790页

作者： Weilin YUAN Jiaxing CHEN Shaofei CHEN Dawei FENG Zhenzhen HU Peng LI Weiwei ZHAO College of Information and Communication National University of Defense TechnologyWuhan 430014China College of Intelligence Science and Technology National University of Defense TechnologyChangsha 410072China Laboratory for Parallel and Distributed Processing National University of Defense TechnologyChangsha 410072China

Reinforcement learning(RL)has become a dominant decision-making paradigm and has achieved notable success in many real-world ***,deep neural networks play a crucial role in unlocking RL’s potential in large-scale decision-making *** by current major success of Transformer in natural language processing and computer vision,numerous bottlenecks have been overcome by combining Transformer with RL for *** paper presents a multiangle systematic survey of various Transformer-based RL(TransRL)models applied in decision-making tasks,including basic models,advanced algorithms,representative implementation instances,typical applications,and known *** work aims to provide insights into problems that inherently arise with the current RL approaches,and examines how we can address them with better TransRL *** our knowledge,we are the first to present a comprehensive review of the recent Transformer research developments in RL for *** hope that this survey provides a comprehensive review of TransRL models and inspires the RL community in its pursuit of future *** keep track of the rapid TransRL developments in the decision-making domains,we summarize the latest papers and their open-source implementations at https://***/williamyuanv0/Transformer-in-Reinforcement-Learning-for-Decision-Making-A-Survey.

关键词： Transformer Reinforcement learning(RL) Decision-making(DM) Deep neural network(DNN) Multi-agent reinforcement learning(MARL) Meta-reinforcement learning(Meta-RL)

来源：评论

学校读者我要写书评

暂无评论

FedEAE: Federated Learning Based Privacy-Preserving Event Argument Extraction 12th

FedEAE: Federated Learning Based Privacy-Preserving Event Ar...

引用

12th National CCF Conference on Natural Language processing and Chinese Computing, NLPCC 2023

作者： Hu, Fei Dong, Shenpo Chang, Tao Zhou, Jie Li, Haili Wang, Jingnan Chen, Rui Liu, Haijiao Wang, Xiaodong National Key Laboratory of Parallel and Distributed Computing College of Computer Science and Technology National University of Defense Technology Changsha410073 China

ISBN: (纸本)9783031446955

Benefiting from Pre-trained Language Model (PLM), Event Argument Extraction (EAE) methods have achieved SOTA performance in general scenarios of Event Extraction (EE). However, with increasing concerns and regulations on data privacy, aggregating distributed data among different institutions in some privacy-sensitive territories (e.g., medical record analysis, financial statement analysis, etc.) becomes very difficult, and it’s hard to train an accurate EAE model with limited local data. Federated Learning (FL) provides promising methods for a large number of clients to collaboratively learn a shared global model without the need to exchange privacy-sensitive data. Therefore, we propose a privacy-preserving EAE method named FedEAE based on FL to solve the current difficulties. To better adapt to federated scenarios, we design a dataset named FedACE generated from the ACE2005 dataset under IID and Non-IID for our experiments. Extensive experiments show that FedEAE achieves promising performance compared to existing baselines, thus validates the effectiveness of our method. To the best of our knowledge, FedEAE is the first to apply FL in the EAE task. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023.

关键词： Extraction

来源：评论

学校读者我要写书评

暂无评论

Deep Time Series Anomaly Detection with Local Temporal Pattern Learning

Deep Time Series Anomaly Detection with Local Temporal Patte...

引用

2025 IEEE International Conference on Acoustics, Speech, and Signal processing, ICASSP 2025

作者： Li, Yizhou Wang, Yijie Xu, Hongzuo Zhou, Xiaohui National Key Laboratory of Parallel and Distributed Computing College of Computer Science and Technology National University of Defense Technology Changsha410073 China Beijing100091 China

ISBN: (纸本)9798350368741

Self-supervised time series anomaly detection (TSAD) demonstrates remarkable performance improvement by extracting high-level data semantics through proxy tasks. Nonetheless, most existing self-supervised TSAD techniques rely on manual- or neural-based transformations when designing proxy tasks, overlooking the intrinsic temporal patterns of time series. This paper proposes a local temporal pattern learning-based time series anomaly detection (LTPAD). LTPAD first generates sub-sequences. Pairwise sub-sequences naturally manifest proximity relationships along the time axis, and such correlations can be used to construct supervision and train neural networks to facilitate the learning of temporal patterns. Time intervals between two sub-sequences serve as labels for sub-sequence pairs. By classifying these labeled data pairs, our model captures the local temporal patterns of time series, thereby modeling the temporal pattern-aware "normality". Abnormal scores of testing data are acquired by evaluating their conformity to these learned patterns shared in training data. Extensive experiments show that LTPAD significantly outperforms state-of-the-art competitors. © 2025 IEEE.

关键词： Local Temporal Pattern Self-supervised Learning Time Series Anomaly Detection

来源：评论

学校读者我要写书评

暂无评论

Adaptive Self-Supervised Continual Learning 26

Adaptive Self-Supervised Continual Learning

引用

26th European Conference on Artificial Intelligence, ECAI 2023

作者： Wu, Lilei Wang, Zhen Liu, Jie Science and Technology on Parallel and Distributed Processing Laboratory National University of Defense Technology Changsha410073 China Laboratory of Digitizing Software for Frontier Equipment National University of Defense Technology Changsha410073 China Tsinghua University Beijing100084 China

ISBN: (纸本)9781643684369

Continual Learning (CL) studies the problem of developing a robust model that can learn new tasks while retaining previously learned knowledge. However, the current CL methods exclusively focus on data with annotations, disregarding that unlabelled data is the mainstream in real-world applications. To close this research gap, this study concentrates on continual self-supervised learning, which is plagued by challenges of memory over-fitting and class imbalance. Besides, these challenges are exacerbated throughout incremental training. Aimed at addressing these challenges from both loss and data perspectives, we introduce a framework, Adaptive Self-supervised Continual Learning (ASCL). Specifically, we devise an Adaptive Sharpness-Aware Minimization (ASAM) module responsible for identifying flatter local minima in the loss landscape with a smaller memory over-fitting risk. Additionally, we design an Adaptive Memory Enhancement (AME) module responsible for rebalancing self-supervised loss with new and old tasks from a data perspective. Finally, the adaptive mechanisms in AME and ASAM modules dynamically adjust the loss landscape sharpness and memory enhancement strength with the feedback of intermediate training results. The results of our extensive experiments demonstrate the state-of-the-art performance of our methods in continual self-supervised learning scenarios across multiple datasets. © 2023 The Authors.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Searching Latent Sub-Goals in Hierarchical Reinforcement Learning as Riemannian Manifold Optimization

Searching Latent Sub-Goals in Hierarchical Reinforcement Lea...

引用

2022 IEEE International Conference on Multimedia and Expo, ICME 2022

作者： Liu, Sidun Qiao, Peng Dou, Yong Jin, Ruochun National University of Defense Technology Science and Technology on Parallel and Distributed Laboratory Hunan China

ISBN: (数字)9781665485630

ISBN: (纸本)9781665485630

Hierarchical Reinforcement Learning (HRL) is promising to tackle the long-term sparse reward problem. However, goal conditioned HRL, which decomposes the goal into a series of sub-goals, suffers from sub-goal search inefficiency problems when the observation space is too large. This problem is more severe in a visual observation space, since its high latent dimensions, where the complete dynamics information is preserved, exponentially increase the difficulty of sub-goal search. In view of this, we propose to treat the latent space as a manifold, i.e., a Riemannian manifold. Assisted by the Riemannian manifold optimization, sub-goals can be efficiently searched in the higher-dimensional latent space, with the help of preserving the dynamics information efficiently. Experiments on a series of MuJoCo tasks with visual observation show that the proposed Riemannian manifold optimization, compared with the baseline that directly searches for sub-goals in bounded latent space, improves the success rate by 1.5 times on average. In much higher dimensions where the baseline no longer converges, the success rate of the proposed method is maintained. © 2022 IEEE.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：