检索结果-内蒙古大学图书馆

Factored Latent-Dynamic conditional random fields for single and multi-label sequence modeling

PATTERN RECOGNITION 2022年 122卷 108236-108236页

作者： Neogi, Satyajit Dauwels, Justin Nanyang Technol Univ Sch Elect & Elect Engn 50 Nanyang Ave Singapore 639798 Singapore

conditional random fields (CRF) are frequently applied for labeling and segmenting sequence data. Morency et al. (2007) introduced hidden state variables in a labeled CRF structure in order to model the latent dynamics within class labels, thus improving the labeling performance. Such a model is known as Latent-Dynamic CRF (LDCRF). We present Factored LDCRF (FLDCRF), a structure that allows multiple latent dynamics of the class labels to interact with each other. Including such latent-dynamic interactions leads to improved labeling performance on single-label and multi-label sequence modeling experiments across two different datasets, viz., UCI gesture phase data and UCI opportunity data. FLDCRF outperforms all state-of-the-art sequence models, viz., CRF, LDCRF, L STM, L STM-CRF, Factorial CRF, Coupled CRF and a multi-label LSTM model across experiments in this paper. In addition, FLDCRF offers easier model selection and is more consistent across validation and test data than LSTM models. FLDCRF is also much faster to train compared to LSTM, even without a GPU. FLDCRF outshines the best LSTM model by-4% on a single-label task on the UCI gesture phase data and outperforms LSTM models by-2% on average on the multi-label sequence tagging experiment on the UCI opportunity data. (c) 2021 Elsevier Ltd. All rights reserved.

关键词： conditional random fields Sequence labeling Multi-task learning Latent-Dynamic models Probabilistic graphical models

来源：评论

学校读者我要写书评

暂无评论

Relevance Propagation through Deep conditional random fields 48

Relevance Propagation through Deep Conditional Random Fields

引用

48th IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2023

作者： Yang, Xiangyu Joukovsky, Boris Deligiannis, Nikos Etro Department Pleinlaan 2 BrusselsB-1050 Belgium Imec Kapeldreef 75 LeuvenB-3001 Belgium

ISBN: (纸本)9781728163277

conditional random fields (CRFs), a particular type of graph neural networks (GNNs), can be used to make structured predictions in machine learning, with various applications from image processing and natural language processing to recommender systems. CRFs refine the prediction of a sample by taking into account its context information. However, there is a lack of work on post-hoc explanation approaches to CRFs, especially when the model is softmax-activated like the deep mean field network (DMFN). In this paper, we bridge this gap by proposing a layer-wise relevance propagation (LRP) method based on deep Taylor decomposition to explain CRFs, especially the DMFN model. The method considers the intermediate softmax activation layers in DMFN. We use two evaluation settings: top K% deletion and insertion to evaluate the method. Experimental studies on fake news detection using the DMFN model prove the effectiveness of our explanation method compared to the other baseline methods. © 2023 IEEE.

关键词： conditional random fields Explainable AI layer-wise relevance propagation post-hoc explanations

来源：评论

学校读者我要写书评

暂无评论

Inference of activity patterns from urban sensing data using conditional random fields

引用

ENVIRONMENT AND PLANNING B-URBAN ANALYTICS AND CITY SCIENCE 2022年第2期49卷 549-565页

作者： Zhu, Yi Shanghai Univ Finance & Econ Sch Publ Econ & Adm Shanghai Peoples R China

The study of human activity-travel patterns for urban planning has evolved a long way in theories, methodologies, and applications. However, the scarcity of data has become a major barrier for the advancement of research in the field. Recently, the proliferation of urban sensing and location-based devices generates voluminous streams of spatio-temporal registered information. In this study, we propose an approach using the linear-chain conditional random fields (CRFs) model to learn the spatio-temporal correspondences of different types of activities and the inter-dependencies among sequential activities from training dataset such as the household travel or time use surveys, and to infer the hidden activity types associated with urban sensing data. The performance of the CRFs model is compared against the random Forest (RF) model, which has been used in a number of existing studies. The results show that the linear-chain CRFs models generally outperform the RF counterparts with respect to classification accuracy of activity types, in particular for those travelers having more outdoor daily activities. The proposed methodology is demonstrated by reconstructing the activity landscape of the surrounding area of a major Mass Rail Transit station in Singapore using the transit smart card transaction data. The inferred activities from the transit smart card data are expected to complement the ground surveys and improve our understanding of the interactions of different components of activities/travels as well as the relationship between urban space and human activities.

关键词： Urban sensing activity-based research activity classification urban dynamics conditional random fields

来源：评论

学校读者我要写书评

暂无评论

Context Model for Pedestrian Intention Prediction Using Factored Latent-Dynamic conditional random fields

引用

IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS 2021年第11期22卷 6821-6832页

作者： Neogi, Satyajit Hoy, Michael Dang, Kang Yu, Hang Dauwels, Justin Nanyang Technol Univ Elect & Elect Engn Dept Singapore 639798 Singapore

Smooth handling of pedestrian interactions is a key requirement for Autonomous Vehicles (AV) and Advanced Driver Assistance Systems (ADAS). Such systems call for early and accurate prediction of a pedestrian's crossing/not-crossing behaviour in front of the vehicle. Existing approaches to pedestrian behaviour prediction make use of pedestrian motion, his/her location in a scene and static context variables such as traffic lights, zebra crossings etc. We stress on the necessity of early prediction for smooth operation of such systems. We introduce the influence of vehicle interactions on pedestrian intention for this purpose. In this paper, we show a discernible advance in prediction time aided by the inclusion of such vehicle interaction context. We apply our methods to two different datasets, one in-house collected - NTU dataset and another public real-life benchmark - JAAD dataset. We also propose a generalization of the Latent-Dynamic conditional random fields (LDCRF), called Factored LDCRF (FLDCRF), for improved sequence prediction performance. FLDCRF outperforms Long Short-Term Memory (LSTM) networks across the datasets over identical time-series features. While the existing best system predicts pedestrian stopping behaviour with 70% accuracy 0.38 seconds before the actual events, our system achieves such accuracy at least 0.9 seconds on an average before the actual events across datasets.

关键词： Hidden Markov models Predictive models Dynamics Vehicle dynamics Context modeling Safety Task analysis Autonomous vehicles intention prediction context models probabilistic graphical models conditional random fields

来源：评论

学校读者我要写书评

暂无评论

An improved gaussian mixture hidden conditional random fields model for audio-based emotions classification

引用

EGYPTIAN INFORMATICS JOURNAL 2021年第1期22卷 45-51页

作者： Siddiqi, Muhammad Hameed Jouf Univ Dept Comp Sci Sakaka Saudi Arabia

The analysis of human emotions plays a significant role in providing sufficient information about patients in monitoring their feelings for better management of their diseases. Audio-based emotions recognition has become a fascinating research interest for such domains during the last decade. Mostly, audio-based emotions systems depend on the recognition stage. The existing model has a common issue called objectivity suppositions problem, which might decrease the recognition rate. Therefore, this study investigates the improved version of a classifier that is based on hidden conditional random fields (HCRFs) model to classify emotional speech. In this model, we introduced a novel methodology that will incorporate multifaceted dissemination with the help of employing a combination of complete covariance Gaussian concreteness function. Due to this incorporation, the proposed model tackle most of the limitations of existing classifiers. Some of the well-known features like Mel-frequency cepstral coefficients (MFCC) are extracted in our experiments. The proposed model has been validated and evaluated on two publicly available datasets likes Berlin Database of Emotional Speech (Emo-DB) and the eNTER FACE'05 Audio-Visual Emotion dataset. For validation and comparison against the existing techniques, we utilized 10-fold cross validation scheme. The proposed method achieved significant improvement under the p-value <0.03 for classification. Moreover, we also prove that computational wise, our computation technique is less expensive against state of the art works. (C) 2020 THE AUTHORS. Published by Elsevier BV on behalf of Faculty of Computers and Artificial Intelligence, Cairo University.

关键词： Emotion classification conditional random fields Hidden markov model Gaussian mixture model

来源：评论

学校读者我要写书评

暂无评论

Parametric geological model update and probabilistic analysis of shield tunnel excavation: A borehole-based conditional random fields approach

引用

COMPUTERS AND GEOTECHNICS 2023年第1期157卷

作者： Xie, Ping Chen, Ke Skibniewskai, Miroslaw Jan Wang, Jun Luo, Hanbin Huazhong Univ Sci & Technol Natl Ctr Technol Innovat Digital Construct Wuhan Peoples R China Huazhong Univ Sci & Technol Sch Civil & Hydraul Engn Wuhan Peoples R China Univ Maryland Ctr Excellence Project Management College Pk MD USA China Railway 14th Bur Grp Mega Shield Constructio Nanjing Peoples R China

Shield tunnel construction has faced considerable challenges due to the uncertain geological conditions in real scenarios. The borehole data obtained from geological investigations are often too limited to effectively evaluate uncertainties at nonsampled points. Moreover, the existing probabilistic safety assessment (PSA) frameworks that consider the uncertainties of soil properties have encountered limitations in engineering practice due to the complexity of conditional random field (CRF) theory and the operability of numerical modeling. Based on the digitization of the underground infrastructure, this study implements the object-oriented parametric modeling (PM) to address these limitations. A PM-CRF engine is first developed to automate the process of parametric geological modeling using CRFs. Then, a PM-CRF-PSA interoperability framework is proposed to automatically generate a PSA model for comprehensive safety assessment considering the uncertainties of soil properties. Finally, a case study is conducted with the proposed framework to assess the whole process of parametric modeling from CRF characterization to PSA. The results of this study highlight the clear advantages of this framework compared to traditional deterministic safety assessment (DSA) methods in terms of characterization of the uncertainty of soil properties and the efficiency of PSA modeling. Thus, the proposed framework enables the efficient and comprehensive exploration of shield tunnel design schemes by considering the uncertainties of soil properties.

关键词： Shield tunnel Parametric modeling conditional random fields Probabilistic analysis Geological conditions

来源：评论

学校读者我要写书评

暂无评论

3D conditional random fields simulation for rockfill compaction quality assessment with sparse EVD measurement

引用

RESULTS IN ENGINEERING 2024年 21卷

作者： Zhao, Lu Cheng, Lili Zhou, Cheng Ding, Lieyun Wang, Fan CCCC Wuhan Zhi Xing Int Engn Consulting Co Ltd Wuhan Peoples R China Huazhong Univ Sci & Technol Natl Ctr Technol Innovat Digital Construct Wuhan 430074 Peoples R China Huazhong Univ Sci & Technol Sch Civil & Hydraul Engn Wuhan 430074 Peoples R China

The compaction quality of rockfill materials plays a pivotal role in ensuring the stability and durability of runway structures. However, traditional assessment methods relying on limited random sampling may lack representativeness for the entire construction site and overlook potential areas of insufficient compaction. Addressing this challenge, this paper employs the dynamic deformation modulus (EVD), a rapid and nondestructive testing method, to assess the compaction quality of runway rockfill. Subsequently, this paper propose a 3D conditional random Field (CRF) model based on EVD to simulate the uncertain spatial distribution of rockfill compaction quality, allowing for the estimation of the overall construction area's compaction quality with a limited number of sample points. The validity of the model is demonstrated through a case study involving a runway expansion project. Additionally, three critical construction factors-namely, measurement points, compaction passes, and compaction materials-are analyzed throughout the construction process. This analysis provides managers with valuable insights and decision-making tools to promptly evaluate the quality of compaction areas, ensuring the stability and durability of the runway.

关键词： Runway Compaction quality assessment EVD Spatial distribution conditional random fields

来源：评论

学校读者我要写书评

暂无评论

conditional random fields Based on Weighted Feature Difference Potential for Remote Sensing Image Classification

Conditional Random Fields Based on Weighted Feature Differen...

引用

Future of Information and Communication Conference (FICC)

作者： Sun, Yi Tian, Yan Xu, Yiping Huazhong Univ Sci & Technol Sch Elect Informat & Commun Natl Key Lab Sci & Technol Multispectral Informat Wuhan 430074 Peoples R China

ISBN: (纸本)9783030394424;9783030394417

Image pixel-wise classification is an important challenge in the field of remote sensing. Some traditional methods, e.g., SVM and random forest, tend to ignore the contextual information, which may fail to process ambiguous samples. conditional random fields (CRF) is an effective approach which eliminates semantic ambiguity via considering the relationship between adjacent pixels. In CRF model, the pairwise potential reflects the contextual relationship. However, the existed pairwise potential often equally treats each feature and does not highlight the roles of key features. In this work, we propose a CRF model with weighted feature difference pairwise potential which makes the key features play important roles. Thus, we can meticulously depict the contextual relationship. We further employ a max-margin method to optimize the CRF model. Experimental results on the hyper-spectral and multi-spectral remote sensing image datasets demonstrate that our approach has the better classification performance.

关键词： Remote sensing Image classification conditional random fields Weighted feature difference potential

来源：评论

学校读者我要写书评

暂无评论

Radar Phrase Extraction Based on Deep Convolutional conditional random fields

Radar Phrase Extraction Based on Deep Convolutional Conditio...

引用

Robotics, Intelligent Control and Artificial Intelligence (RICAI), International Conference on

作者： Wenjing He Yang-Yang Dong Lidong Zhang Chunxi Dong School of Electronic Engineering Xidian University Xi’an China Unit 93209 of the PLA Beijing China

ISBN: (数字)9798331541699

ISBN: (纸本)9798331541705

Extracting radar phrases from multifunctional radar pulse sequences and identifying the internal operation law and working mechanism among the working modes is the foundation for subsequent working mode prediction. However, most existing studies at the syntactic level fail to address the full range of issues, focusing only on radar word extraction and working pattern recognition. In this paper, we propose an improved CNN-BiLSTM-CRF model for radar phrase extraction. It is based on the BiLSTM network model, which is able to extract richer local features using convolutional neural networks. The model is simulated on generated radar pulse word sequences, and the results show that the CNN-BiLSTM-CRF model improves the F-value of radar phrase extraction by 2.688% on the test set compared with the BiLSTM-CRF model. The simulation results prove that the improved method is effective.

关键词： Training Simulation Semantics Bidirectional long short term memory Radar Syntactics Feature extraction conditional random fields Convolutional neural networks Robots

来源：评论

学校读者我要写书评

暂无评论

Human Motion Trajectory Perception and Recognition Based on Multi Sensor Network Fusion with conditional random fields

Human Motion Trajectory Perception and Recognition Based on ...

引用

Mobile Networks and Wireless Communications (ICMNWC), IEEE International Conference on

作者： Yuanming Jiu Qimeng Niu Jiacheng Feng Department of Physical Education Tianjin University of Technology Tianjin China

ISBN: (数字)9798331528348

ISBN: (纸本)9798331528355

In recent years, effective modeling and capturing realistic human motion in 3Dimensional (3D) environments was crucial for applications including game character animation, virtual reality and human-robot interaction. Traditional approaches for human motion recognition had faced several challenges which include limited generalizability, noise and occlusion, variability in motion patterns. Therefore, this research proposes Multi Sensor Network Fusion with conditional random fields (MSNFCRF) to recognize human motion trajectory perception. Initially, the data is taken from Human3.6M dataset which consists of human pose data and motion capture data. Then, this data is preprocessed by using Empirical Mode Decomposition (EMD) which effectively reduced noise and dimensionality. After that, the features are extracted by using Maximum Lyapunor Exponent (MLE) which extracted dynamic and chaotic behavior of human motions and identifies subtle patterns and variations in the motion data. Finally, recognition is done by using proposed MSNFCRF which enhanced situational awareness and better handling of complex scenarios and effectively integrates information from multiple scales and modalities by enabling it to capture subtle patterns and relationships in the data. The proposed MSNFCRF achieved better results of accuracy (98.91 %), precision (98.83 %), recall (98.82%) and F1-score (98.90%) when compared with existing Inception-Long Short-Term Memory.

关键词： Wireless communication Maximum likelihood estimation Accuracy Noise Dynamics Feature extraction Motion capture conditional random fields Trajectory Data mining

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：