检索结果-内蒙古大学图书馆

20th OITS International Conference on Information Technology, OCIT 2022

作者： Srivastava, Smita Gupta, Deepa Paul, Biswajit Sahoo, Shubhashisa Amrita School of Computing Amrita Vishwa Vidyapeetham Dept of Computer Science and Engineering Bangalore India Défense Research and Development Organization Centre of Artificial Intelligence and Robotics Bangalore India

ISBN: (纸本)9781665493482

A vast majority of cybersecurity data comes in the form of unstructured textual data and needs to be annotated proficiently to train supervised machine learning models. The critical question is how much and which subset of data should be annotated for better model performance under budget constraints. Though most of the Machine Learning (ML) research focuses on learning better models using annotated datasets, this paper focuses on data annotation, specifically on suitable subset selection with an emphasis on Named Entity Recognition (NER) for cybersecurity. The proposed method provides an active learning based sampling strategy to select minimal yet most informative samples from a large set. Further, reinforcement learning is combined with the active learning approach to automate the process of sampling. The results on the auto-labelled cyber-NER dataset indicate that the cyber-NER model with Reinforced Active Learning (RAL) based sampling increases F1-Score by +2-7% and reduces compute time by 90% compared to random sampling based subset selection. Further, the proposed RAL approach achieved an 80% reduction in sample size and, consequently, annotation cost with comparable accuracy to that of complete selection. © 2022 IEEE.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

A Discontinuous Galerkin Time-Domain Scheme to Model Lasing Dynamics in Four-Level Two-Electron Atomic Systems

A Discontinuous Galerkin Time-Domain Scheme to Model Lasing ...

引用

European Conference on Antennas and Propagation, EuCAP

作者： Ming Dong Liang Chen Ran Zhao Hakan Bagci Electrical and Computer Engineering Program Computer Electrical and Mathematical Science and Engineering Division King Abdullah University of Science and Technology (KAUST) Thuwal Saudi Arabia School of Electronic Science and Engineering Univesirty of Electronic Science and Technology of China (UESTC) Chengdu China

A discontinuous Galerkin time-domain scheme is formulated and implemented to analyze three-dimensional transient lasing dynamics. The proposed scheme solves a coupled system of the Maxwell and the rate equations. The atomic transitions through different energy levels are quantum-mechanically described by a four-level two-electron model, while the electro-magnetic interactions are treated using the Maxwell equations. The resulting solver accounts for the Pauli Exclusion Principle and permits robust simulation of lasing dynamics under optical pumping. Numerical results are presented to demonstrate the applicability and the accuracy of the proposed scheme.

关键词：

来源：评论

学校读者我要写书评

暂无评论

An Perception Enhanced Human-Robot Skill Transfer Method for Reactive Interaction 25th

An Perception Enhanced Human-Robot Skill Transfer Method for...

引用

25th Annual Conference on Towards Autonomous Robotic Systems, TAROS 2024

作者： Si, Weiyong Dong, Jiale Wang, Ning Yang, Chenguang School of Computer Science and Electronic Engineering University of Essex Wivenhoe Park ColchesterCO4 3SQ United Kingdom South China University of Technology Guangzhou510640 China Bristol Robotics Lab University of the West of England BristolBS16 1QY United Kingdom Department of Computer Science University of Liverpool LiverpoolL69 3BX United Kingdom

ISBN: (纸本)9783031720611

In this paper, we propose a trust human-robot skill transfer framework, interpretive and reactive dynamic system (IRDS), by investigating the behaviour tree (BT) and dynamic movement primitives (DMPs) enhanced by the feedback from perceived information for dynamic and uncertain tasks and environments. Human sensorimotor control allows interactions with various environments and accomplishes complex manipulation tasks with uncertainty;however, it is still hard for robots to own this capability. In this work, we aim to transfer these reactive skills to robots, which enable robots to interact with humans and environments under varying uncertainty. The main challenges of robot skill learning are the generalization, safety and stability issues during the skills learning and execution autonomously. BT was investigated for task planning and reactive interaction during the robot execution. The dynamic system-based model generates action for the low-level compliant controller. The convergence of the proposed IRDS framework under dynamic interaction and disturbance was proved by control theory. We conducted simulations and physical experiments on the real robots to evaluate the generalization performance and the convergence capability under uncertainty. And the results show that the reactivity, convergence and interaction performance can be guaranteed, and the learned skills can be transferred among different physical robot platforms. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Human robot interaction

来源：评论

学校读者我要写书评

暂无评论

Study of Word Embeddings for Enhanced Cyber Security Named Entity Recognition

引用

Procedia computer science 2023年 218卷 449-460页

作者： Smita Srivastava Biswajit Paul Deepa Gupta Dept of Computer Science and Engineering Amrita School of Engineering Bengaluru Amrita Vishwa Vidyapeetham India Centre for Artificial Intelligence and Robotics Défense Research and Development Organization Bengaluru India

A vast majority of cyber security information is in the form of unstructured text. A much-needed task is to have a machine-assisted analysis of such information. Named Entity Recognition (NER) provides a vital step towards this conversion. However, cyber security named entities are not restricted to classical entity types like people, location, organisation, miscellaneous etc but comprise a large set of domain-specific entities. Word embedding has emerged as the dominant choice for the initial transfer of semantics to downstream NLP tasks and impacts performance. Though several word embeddings learned using general purpose large corpus like Google News, Wikipedia etc. are available as pre-trained embeddings and have shown good performance on NER tasks; this trend is not consistent when it comes to domain-specific NER. This work explores the relative performances and suitability of prominent word embeddings for cyber security NER task. Embeddings considered include both general-purpose pre-trained word embeddings (non-contextual and contextual) available in the public domain and task-adapted embedding generated by fine-tuning these pre-trained embeddings on a task-specific supervised dataset. The results indicate that when it comes to using pre-trained embeddings for cyber security NER, fastText performs better than GloVe and BERT. However, when embeddings are further fine-tuned for the cyber-NER task, the performance of all the fine-tuned embeddings improved by +2-7%. Further, BERT embedding fine-tuned using position-wise FFN (Feed Forward Network) produced the state-of-the-art 0.974 F1-Score on the cyber security NER dataset.

关键词： Cyber Security Named Entity Recognition Word Embeddings BiLSTM CRF fastText GloVe BERT Fine-Tuning

来源：评论

学校读者我要写书评

暂无评论

Breast Tumor Classification with Fine-Tuned Hyperparameter Training using Deep Learning Models

Breast Tumor Classification with Fine-Tuned Hyperparameter T...

引用

AI-Driven Smart Healthcare for Society 5.0

作者： Yash Maini Sanjay Kumar Singh Pranshu Saxena University School of Automation and Robotics Guru Gobind Singh Indraprastha University Surajmal Vihar Delhi India School of Computer Science Engineering and Technology Bennett University Greater Noida India

ISBN: (数字)9798331536336

ISBN: (纸本)9798331536343

Breast tumor classification plays a crucial role in early detection and treatment planning for breast cancer. In recent years, deep learning models have shown remarkable success in image-based medical diagnosis. This study focuses on improving breast tumor classification using fine-tuned hyperparameter training of deep learning models. By optimizing key parameters such as learning rate, batch size, and network architecture, the classification accuracy of models like ResNet, AlexNet, GoogleNet and VGG is enhanced. The proposed approach leverages advanced techniques such as transfer learning, data augmentation, and hyperparameter optimization algorithms (Bayesian Optimization along with population-based training) to achieve robust performance on breast cancer datasets. The proposed method is able to achieve 98.46% accuracy while classifying breast cancer dataset. Experimental results demonstrate significant improvements in classification accuracy, precision, recall, and F1-score compared to conventional approaches, making the fine-tuned models more suitable for clinical applications. The findings suggest that hyperparameter tuning in deep learning models is critical to improving model performance and generalization in breast tumor classification tasks.

关键词： Deep learning Training Accuracy Breast tumors Ultrasonic imaging Transfer learning Breast cancer Planning Optimization Tuning

来源：评论

学校读者我要写书评

暂无评论

SEQUEL: Semi-Supervised Preference-based RL with Query Synthesis via Latent Interpolation

SEQUEL: Semi-Supervised Preference-based RL with Query Synth...

引用

IEEE International Conference on robotics and Automation (ICRA)

作者： Daniel Marta Simon Holk Christian Pek Iolanda Leite Division of Robotics Perception and Learning School of Electrical Engineering and Computer Science KTH Royal Institute of Technology Stockholm Sweden Digital Futures Dept. of Cognitive Robotics TU Delft Delft The Netherlands

ISBN: (数字)9798350384574

ISBN: (纸本)9798350384581

Preference-based reinforcement learning (RL) poses as a recent research direction in robot learning, by allowing humans to teach robots through preferences on pairs of desired behaviours. Nonetheless, to obtain realistic robot policies, an arbitrarily large number of queries is required to be answered by humans. In this work, we approach the sample-efficiency challenge by presenting a technique which synthesizes queries, in a semi-supervised learning perspective. To achieve this, we leverage latent variational autoencoder (VAE) representations of trajectory segments (sequences of state-action pairs). Our approach manages to produce queries which are closely aligned with those labeled by humans, while avoiding excessive uncertainty according to the human preference predictions as determined by reward estimations. Additionally, by introducing variation without deviating from the original human’s intents, more robust reward function representations are achieved. We compare our approach to recent state-of-the-art preference-based RL semi-supervised learning techniques. Our experimental findings reveal that we can enhance the generalization of the estimated reward function without requiring additional human intervention. Lastly, to confirm the practical applicability of our approach, we conduct experiments involving actual human users in a simulated social navigation setting. Videos of the experiments can be found at https://***/view/rl-sequel

关键词： Interpolation Uncertainty Navigation Estimation Reinforcement learning Semisupervised learning Robot learning

来源：评论

学校读者我要写书评

暂无评论

Unsupervised Cross-Domain 3D Human Pose Estimation via Pseudo-Label-Guided Global Transforms

arXiv

引用

arXiv 2025年

作者： Liu, Jingjing Wang, Zhiyong Fan, Xinyu Dadashzadeh, Amirhossein Liu, Honghai Mirmehdi, Majid School of Computer Science University of Bristol United Kingdom State Key Laboratory of Robotics and Systems Harbin Institute of Technology Shenzhen China School of Aerospace Engineering Xiamen University China

Existing 3D human pose estimation methods often suffer in performance, when applied to cross-scenario inference, due to domain shifts in characteristics such as camera viewpoint, position, posture, and body size. Among these factors, camera viewpoints and locations have been shown to contribute significantly to the domain gap by influencing the global positions of human poses. To address this, we propose a novel framework that explicitly conducts global transformations between pose positions in the camera coordinate systems of source and target domains. We start with a Pseudo-Label Generation Module that is applied to the 2D poses of the target dataset to generate pseudo-3D poses. Then, a Global Transformation Module leverages a human-centered coordinate system as a novel bridging mechanism to seamlessly align the positional orientations of poses across disparate domains, ensuring consistent spatial referencing. To further enhance generalization, a Pose Augmentor is incorporated to address variations in human posture and body size. This process is iterative, allowing refined pseudo-labels to progressively improve guidance for domain adaptation. Our method is evaluated on various cross-dataset benchmarks, including Human3.6M, MPI-INF-3DHP, and 3DPW. The proposed method outperforms state-of-the-art approaches and even outperforms the target-trained model. Copyright © 2025, The Authors. All rights reserved.

关键词： Three dimensional computer graphics

来源：评论

学校读者我要写书评

暂无评论

Generalizing 6-DoF Grasp Detection via Domain Prior Knowledge

Generalizing 6-DoF Grasp Detection via Domain Prior Knowledg...

引用

Conference on computer Vision and Pattern Recognition (CVPR)

作者： Haoxiang Ma Modi Shi Boyang Gao Di Huang State Key Laboratory of Software Development Environment Beihang University Beijing China School of Computer Science and Engineering Beihang University Beijing China School of Computer Science and Technology Harbin Institute of Technology Harbin China Geometry Robotics

ISBN: (数字)9798350353006

ISBN: (纸本)9798350353013

We focus on the generalization ability of the 6-DoF grasp detection method in this paper. While learning-based grasp detection methods can predict grasp poses for unseen ob-jects using the grasp distribution learned from the training set, they often exhibit a significant performance drop when encountering objects with diverse shapes and struc-tures. To enhance the grasp detection methods' general-ization ability, we incorporate domain prior knowledge of robotic grasping, enabling better adaptation to objects with significant shape and structure differences. More specifi-cally, we employ the physical constraint regularization during the training phase to guide the model towards predicting grasps that comply with the physical rule on grasping. For the unstable grasp poses predicted on novel objects, we design a contact-score joint optimization using the pro-jection contact map to refine these poses in cluttered sce-narios. Extensive experiments conducted on the GraspNet-1 billion benchmark demonstrate a substantial performance gain on the novel object set and the real-world grasping experiments also demonstrate the effectiveness of our gen-eralizing 6-DoF grasp detection method. Code is available at https://***/mahaoxiang822/Generalizing-Grasp.

关键词： Training computer vision Shape Grasping Benchmark testing Predictive models Performance gain

来源：评论

学校读者我要写书评

暂无评论

A Novel Noise Filtering Method Using Kalman Filter in Digital Holographic Microscopy (DHM)

A Novel Noise Filtering Method Using Kalman Filter in Digita...

引用

International Technical Conference o Circuits/Systems, computers and Communications (ITC-CSCC)

作者： Taishi Ono Hyun-Woo Kim Myungjin Cho Min-Chul Lee Graduate School of Computer Science and Systems Engineering Kyushu Institute of Technology Fukuoka Japan Department of Computer Science and Networks Kyushu Institute of Technology Fukuoka Japan IITC School of ICT Robotics and Mechanical Engineering Hankyong National University Kyonggi-do Republic of Korea

ISBN: (数字)9798350379051

ISBN: (纸本)9798350379068

Three-dimensional (3D) profile by Digital Holographic Microscopy (DHM) always contains noise. This is an issue in the phase information extraction method of DHM system. However, in the conventional phase information extraction from the frequency domain, it may not extract only phase information. To solve this problem, we propose that the extracting phase information from the frequency domain of recorded hologram using DHM system is specified multiple times and Kalman filtering is performed on the multiple 3D profiles. The purpose of this method is to eliminate randomly generated noise and generate highly accurate 3D profiles. The generated 3D profile is subjectively and quantitatively evaluated with conventional methods. As a result, 3D profile of the proposed method shows the higher accuracy than the conventional method.

关键词： Three-dimensional displays Accuracy Shape Microscopy Frequency-domain analysis Noise Noise reduction

来源：评论

学校读者我要写书评

暂无评论

Nested Deep Feature Attention Module for Underwater Image Enhancement 3

Nested Deep Feature Attention Module for Underwater Image En...

引用

3rd International conference on computer, Big Data and Artificial Intelligence, ICCBDAI 2022

作者： Xiao, Yeqing Wang, Qiang Li, Yupeng Tang, Yandong School of Computer Science and Engineering Shenyang Jianzhu University Shenyang110168 China Key Laboratory of Manufacturing Industrial Integrated Shenyang University Shenyang110044 China State Key Laboratory of Robotics Shenyang Institute of Automation Chinese Academy of Sciences Shenyang110016 China Institutes for Robotics and Intelligent Manufacturing Chinese Academy of Sciences Shenyang110169 China

The captured underwater images always suffer degradations because of absorption and light scattering in water. Thus, underwater image enhancement becomes indispensable as a precondition to carry out underwater tasks. Thus, we put forward an end-to-end structure to solve the problem of underwater image degradation. Our network uses continuously stacked deep-layer feature extraction modules to exploit more significant features. For purpose of enhancing the image contrast, we recover high-frequency information by hiring the nested residual attention groups in feature extraction modules, which can adaptively extract more deeper features. Moreover, channel attention module is hired in residual groups to emphasize important information. Comparing with other state-of-the-art approaches, our network performs best on quantitative and qualitative evaluations at datasets EUVP and UIEBD. © Published under licence by IOP Publishing Ltd.

关键词： Feature extraction

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：