检索结果-内蒙古大学图书馆

IEEE Transactions on Audio, Speech and Language processing 2025年 33卷 2074-2083页

作者： Shoutao Guo Shaolei Zhang Zhengrui Ma Min Zhang Yang Feng Key Laboratory of Intelligent Information Processing Institute of Computing Technology Chinese Academy of Sciences (ICT/CAS) Beijing China University of Chinese Academy of Sciences Beijing China School of Computer Science and Technology Soochow University Suzhou China Key Laboratory of AI Safety Chinese Academy of Sciences Beijing China

Simultaneous Machine Translation (SiMT) generates target translations in real-time while reading the source sentence. It relies on a policy to determine the optimal timing for producing translations, aiming to achieve both low response latency and high translation quality. Existing SiMT methods typically utilize the traditional Transformer architecture, employing complicated dynamic programming algorithms to jointly optimize policy decisions and translation generation. While these methods excel at determining policies, their translation capabilities are suboptimal. In contrast, Large Language Models (LLMs), trained on extensive corpora, exhibit exceptional generation capabilities but struggle to learn translation policies through traditional Transformer-based training methods. To address these limitations, we introduce Agent-SiMT, a novel framework combining the strengths of LLMs and traditional SiMT methods. Agent-SiMT contains a policy-decision agent and a translation agent. The policy-decision agent is managed by a SiMT model, which determines the policy using partial source sentences and translations. The translation agent, powered by an LLM, generates the target translation according to the policy decisions. These two agents work collaboratively to perform SiMT. Experiments demonstrate that our method attains state-of-the-art performance.

关键词： Translation Hidden Markov models Transformers Training Large language models Collaboration Vocabulary Speech processing Memory modules Machine translation

来源：评论

学校读者我要写书评

暂无评论

DATG: Data Augmentation with Transformer-Based Generation for Low-Resource Named Entity Recognition

DATG: Data Augmentation with Transformer-Based Generation fo...

引用

Chinese Automation Congress (CAC)

作者： Qian Yili Xu Haonan Key Laboratory of Computational Intelligence and Chinese Information Processing of Ministry of Education School of Computer and Information Technology Shanxi University Taiyuan China School of Computer and Information Technology Shanxi University Taiyuan China

Data augmentation is an important technique for enhancing machine learning performance. In this study, we propose a novel generative data augmentation method for named entity recognition, which addresses the challenge of limited labeled data in professional domains such as biomedical and chemical. The proposed method involves organizing input entities into a Transformer model to produce tagged sentences that are specifically related to the input entities. By leveraging external entity dictionaries, we effectively generate new tagged data and demonstrate its superiority over other data augmentation methods. Our approach offers a potential solution to the NER challenges faced in low-resource languages and specialized domains.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Tissue P Systems with Look-Ahead Mode

引用

Chinese Journal of Electronics 2025年第1期23卷 81-86页

作者： Yun Jiang Tao Song Zheng Zhang School of Computer Science and Information Engineering Chongqing Technology and Business University Chongqing China Department of Control Science and Engineering Image Processing and Intelligent Control Key Laboratory of Education Ministry of China Huazhong University of Science and Technology Wuhan China

Tissue P systems are a class of distributed and parallel computing models inspired from inter-cellular communication and cooperation between cells. In this work, a variant of tissue P system, named tissue P system with look-ahead mode, is discussed for decreasing the inherent non-determinism of tissue P systems and helping implementing tissue P systems on computers. Such systems are proved to be universal by simulating register machine, and they are also proved to be able to efficiently solve computationally hard problems by means of a space-time tradeoff, which is illustrated with a polynomial solution to 3-coloring problem.

关键词： computers Computational modeling Medical services Parallel processing Polynomials Computational efficiency Registers Business

来源：评论

学校读者我要写书评

暂无评论

MRA-Depth: Multi-Resolution Attention based Monocular Depth Estimation

MRA-Depth: Multi-Resolution Attention based Monocular Depth ...

引用

Robotics and computer Vision (ICRCV), International Conference on

作者： Jing Wen Lukuan Niu School of computer information and technology Key laboratory of computational intelligence and chinese information processing of ministry of education Shanxi University Taiyuan China School of computer information and technology Shanxi University Taiyuan China

Delicate feature would become weaker or even disappear as going through deeper network layer, which results in jagged edges of objects in the depth estimation. To address this problem, we propose a multi-resolution attention based monocular depth estimation by capturing and recalibrating the dependencies of channels between different resolutions, including: we develop a multi-resolution attention encoder by embedding resolution and spatial attention, so as to extract the higher quality features in encoder, and we design a reusable multi-resolution fusion model in decoder, in which different scale features are integrated and then extract multi-resolution attention maps, so as to enhance detail information locally and perspective structure features globally. Experimental results illustrate the proposed network is capable to predict more precise depth and sharper contour, and to outperforms related methods on the KITTI benchmark.

关键词：

来源：评论

学校读者我要写书评

暂无评论

SNIDA: Unlocking Few-Shot Object Detection with Non-Linear Semantic Decoupling Augmentation

SNIDA: Unlocking Few-Shot Object Detection with Non-Linear S...

引用

Conference on computer Vision and Pattern Recognition (CVPR)

作者： Yanjie Wang Xu Zou Luxin Yan Sheng Zhong Jiahuan Zhou Huazhong University of Science and Technology Wuhan China National Key Laboratory of Multispectral Information Intelligent Processing Technology China Wangxuan Institute of Computer Technology Peking University Beijing China

ISBN: (数字)9798350353006

ISBN: (纸本)9798350353013

Once only a few-shot annotated samples are available, the performance of learning-based object detection would be heavily dropped. Many few-shot object detection (FSOD) methods have been proposed to tackle this issue by adopting image-level augmentations in linear manners. Nevertheless, those handcrafted enhancements often suf-fer from limited diversity and lack of semantic awareness, resulting in unsatisfactory performance. To this end, we propose a Semantic-guided Nonlinear Instance-level Data Augmentation method (SNIDA) for FSOD by decoupling the foreground and background to increase their diversities respectively. We design a semantic awareness enhancement strategy to separate objects from backgrounds. Concretely, masks of instances are extracted by an unsupervised semantic segmentation module. Then the diversity of samples would be improved by fusing instances into different backgrounds. Considering the shortcomings of augmenting images in a limited transformation space of existing traditional data augmentation methods, we introduce an object reconstruction enhancement module. The aim of this module is to generate sufficient diversity and nonlinear training data at the instance level through a semantic-guided masked autoencoder. In this way, the potential of data can be fully exploited in various object detection scenarios. Extensive experiments on PASCAL VOC and MS-COCO demonstrate that the proposed method outperforms base-lines by a large margin and achieves new state-of-the-art results under different shot settings.

关键词： computer vision Fuses Semantic segmentation Computational modeling Semantics Training data Object detection

来源：评论

学校读者我要写书评

暂无评论

Multi-stage Objective Function Optimized Hand-Eye Self-calibration of Robot in Autonomous Environment 1

引用

17th International Conference on Bio-Inspired Computing: Theories and Applications, BIC-TA 2022

作者： Liu, Kaibo Zheng, Wangli Min, Huasong Lin, Yunhan School of Computer Science and Technology Wuhan University of Science and Technology Hubei Wuhan430065 China Hubei Province Key Laboratory of Intelligent Information Processing and Real-Time Industrial System Wuhan University of Science and Technology Hubei Wuhan430065 China State Grid Electric Power Research Institute Jiangsu Nanjing211106 China Institute of Robotics and Intelligent Systems Wuhan University of Science and Technology Hubei Wuhan430081 China

ISBN: (数字)9789819915491

ISBN: (纸本)9789819915484

In the application scenario of robot autonomous tasks, the robot needs to be able to complete calibration online and automatically to achieve self-maintenance, which differs from traditional robot hand-eye calibration in that the traditional robot hand-eye calibration requires a dedicated calibration board to assist offline completion. Aiming at the problem that the existing self-calibration methods cannot be optimized as a whole, which leads to low accuracy and instability of the solution, a multi-stage objective function optimization self-calibration algorithm is proposed, which describes the solution of hand-eye self-calibration as a minimization objective function problem involving multiple stages. An optimization method based on the minimization of re-projection error is designed to compensate for the results, which uses an efficient Oriented fast and rotated brief (ORB) feature extraction algorithm and introduces a scoring mechanism to retain more correct matching points in the feature matching stage. Two different types of experiments were designed to validate our method. One is a single camera dataset experiment, which shows that our method is more accurate and robust than the existing self-calibration method;the other is an application platform experiment, which verifies the feasibility and availability of our method. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Calibration

来源：评论

学校读者我要写书评

暂无评论

Zero-shot event-intensity asymmetric stereo via visual prompting from image domain 24

Zero-shot event-intensity asymmetric stereo via visual promp...

引用

Proceedings of the 38th International Conference on Neural information processing Systems

作者： Hanyue Lou Jinxiu Liang Minggui Teng Bin Fan Yong Xu Boxin Shi State Key Laboratory of Multimedia Information Processing and National Engineering Research Center of Visual Technology School of Computer Science Peking University National Key Laboratory of General AI School of Intelligence Science and Technology Peking University School of Computer Science and Engineering South China University of Technology

ISBN: (纸本)9798331314385

Event-intensity asymmetric stereo systems have emerged as a promising approach for robust 3D perception in dynamic and challenging environments by integrating event cameras with frame-based sensors in different views. However, existing methods often suffer from overfitting and poor generalization due to limited dataset sizes and lack of scene diversity in the event domain. To address these issues, we propose a zero-shot framework that utilizes monocular depth estimation and stereo matching models pretrained on diverse image datasets. Our approach introduces a visual prompting technique to align the representations of frames and events, allowing the use of off-the-shelf stereo models without additional training. Furthermore, we introduce a monocular cue-guided disparity refinement module to improve robustness across static and dynamic regions by incorporating monocular depth information from foundation models. Extensive experiments on real-world datasets demonstrate the superior zero-shot evaluation performance and enhanced generalization ability of our method compared to existing approaches.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Ship flame segmentation by fusing DASPP and PE-SCAM attention 3

Ship flame segmentation by fusing DASPP and PE-SCAM attentio...

引用

3rd International Conference on Neural Networks, information and Communication Engineering, NNICE 2023

作者： Zheng, Xuecong Chen, Li Chen, Yaojie School of Computer Science and Technology Wuhan University of Science and Technology Hubei Wuhan430065 China Wuhan Institute of Marine Electric Propulsion Hubei Wuhan430065 China Hubei Province Key Laboratory of Intelligent Information Processing and Real-time Industrial System Hubei Wuhan430065 China

ISBN: (纸本)9798350335972

To solve the problem that the flame area of small targets on ships can not be easily segmented and the flame boundary is obviously affected by other background such as smoke, this experiment improves DeepLabV3+ algorithm from the following three aspects. Firstly, Dense Atrous Spatial Pyramid Pooling (DASPP) was improved to enhance the extraction and utilization of flame boundary and small target flame features. Secondly, in order to avoid the influence of background such as smoke around the flame, the high-level flame features and shallow flame features obtained by DASPP were combined with the improved Attention Mechanism (PE-SCAM) to obtain more context information in spatial and channel dimensions, so as to enhance the processing effect of the model on the flame and its boundaries. Finally, the traditional cross loss function is optimized to reduce the influence caused by the imbalance of positive and negative samples between the flame and the background. Experimental results show that the MIou of the proposed algorithm on the ship flame data set reaches 88.43%, which can realize the accurate segmentation of small target flames and flame boundaries on the basis of ensuring high real-time performance. © 2023 IEEE.

关键词： Ships

来源：评论

学校读者我要写书评

暂无评论

Temporal Feature Aggregation for Efficient 2D Video Grounding

Temporal Feature Aggregation for Efficient 2D Video Groundin...

引用

IEEE International Conference on Multimedia and Expo (ICME)

作者： Mohan Chen Yiren Zhang Jueqi Wei Yuejie Zhang Rui Feng Tao Zhang Shang Gao School of Computer Science Shanghai Key Laboratory of Intelligent Information Processing Shanghai Collaborative Innovation Center of Intelligent Visual Computing Fudan University School of Information Management and Engineering Shanghai Key Laboratory of Financial Information Technology Shanghai University of Finance and Economics School of Information Technology Deakin University

ISBN: (数字)9798350390155

ISBN: (纸本)9798350390162

Video grounding aims to locate the target video moment in an untrimmed video based on a text query. Most existing methods employ 3D CNNs as the video feature extractor, incurring substantial computational costs. Only a few methods use 2D backbones for video feature extraction, and they suffer from diminished accuracy due to the inherent lack of temporal information within 2D features. To address this problem, we propose a novel 2D video grounding method called TFA that improves accuracy while minimizing computational costs. Our approach involves a query-guided temporal feature aggregation module designed to explicitly capture temporal information. We disentangle time intervals of input video frames and prediction spans to reduce computational overhead. Additionally, we introduce deformable attention into the multi-modal encoder for further enhancement. Extensive experiments on two public datasets demonstrate that our method outperforms previous 2D video grounding methods and achieves competitive results with most 3D methods at significantly reduced costs.

关键词： Three-dimensional displays Accuracy Costs Grounding Streaming media Feature extraction Computational efficiency

来源：评论

学校读者我要写书评

暂无评论

Unleashing Fine-Coarse Curve Perception Via Trunk-Branch Perturbation

Unleashing Fine-Coarse Curve Perception Via Trunk-Branch Per...

引用

IEEE International Conference on Image processing

作者： Yunxiang Cao Li Chen Yubo Wang Zhida Feng Xiaoming Liu School of Computer Science and Technology Wuhan University of Science and Technology Hubei Province Key Laboratory of Intelligent Information Processing and Real-time Industrial System Wuhan University of Science and Technology

ISBN: (数字)9798350349399

ISBN: (纸本)9798350349405

Segmenting intricate curve structures like retinal blood vessels, encompassing both fine and coarse details, remains a significant challenge. This work proposes a novel module that divides these complex curve structures into trunks and branches, fuses the input as auxiliary information, and optimizes the breakpoints of different curve parts through losses. In order to balance the redundant information that may lead to model overfitting, a unique feature perturbation strategy is introduced after the backbone decoding process to enhance the model’s robustness to complex curve structure segmentation tasks. Experiments show that this method can effectively distinguish different topological structures of blood vessels and maintain high segmentation accuracy even at blood vessel intersections or breakpoints, which holds immense potential for diverse future applications in image segmentation.

关键词： Measurement Image segmentation Fuses Perturbation methods Blood vessels Retina Robustness

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

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

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

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

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：