35,361 Hits in 5.9 sec

Semi-supervised multi-task learning of structured prediction models for web information extraction

Paramveer S. Dhillon, Sundararajan Sellamanickam, Sathiya Keerthi Selvaraj
2011 Proceedings of the 20th ACM international conference on Information and knowledge management - CIKM '11  
Viewing the problem of building a local model for each website as a task, we learn the models for a collection of sites jointly; thus our method can also be seen as a graph regularization based multi-task  ...  Learning the models jointly with the proposed method is very useful in two ways: (1) learning a local model for a website can be effectively influenced by labeled and unlabeled data from other websites  ...  multi-task learning scenarios.  ... 
doi:10.1145/2063576.2063713 dblp:conf/cikm/DhillonSS11 fatcat:tqzsw2q3zzccpmbdt2rparxo2q

Inferring Brain Dynamics via Multimodal Joint Graph Representation EEG-fMRI [article]

Jalal Mirakhorli
2022 arXiv   pre-print
Throughout, we outline the correlations of several different media in time shifts from one source with graph-based and deep learning methods.  ...  The joint representations of different modalities is a robust model to analyze simultaneously acquired electroencephalography and functional magnetic resonance imaging (EEG-fMRI).  ...  Due to such a high correlation of features from multi-view and improving the learning process on each perspective, in here apply multi-head attention [39] .  ... 
arXiv:2201.08747v1 fatcat:ia72xmqjrraj3h6pjt2iyf372e

A Relational Adaptive Neural Model for Joint Entity and Relation Extraction

Guiduo Duan, Jiayu Miao, Tianxi Huang, Wenlong Luo, Dekun Hu
2021 Frontiers in Neurorobotics  
In the task of entity relation joint extraction, overlapping entities and multi-type relation extraction in overlapping triplets remain a challenging problem.  ...  A relational-adaptive entity relation joint extraction model based on multi-head self-attention and densely connected graph convolution network (which is called MA-DCGCN) is proposed in the paper.  ...  We compare the results with the MultiDecoder. • CopyMTL (Zeng et al., 2020 ) introduces a multi-task learning framework, which solves the problem of extracting only one word in CopeRe by adopting different  ... 
doi:10.3389/fnbot.2021.635492 pmid:33796016 pmcid:PMC8008121 fatcat:gicxhwevdfgodcephdlnw5xbrm

Control Flow Graph Embedding Based on Multi-Instance Decomposition for Bug Localization

Xuan Huo, Ming Li, Zhi-Hua Zhou
from the control flow graph.  ...  In this paper, we propose a novel model named CG-CNN, which is a multi-instance learning framework that enhances the unified features for bug localization by exploiting structural and sequential nature  ...  After processing from language-specific feature extraction layer, the generated features z r from bug reports and z c from source code are then fed into joint feature learning layer, where a fully-connected  ... 
doi:10.1609/aaai.v34i04.5844 fatcat:e5vffrbkdvc7zoqyf5oqmkg76i

SM-SGE: A Self-Supervised Multi-Scale Skeleton Graph Encoding Framework for Person Re-Identification [article]

Haocong Rao, Xiping Hu, Jun Cheng, Bin Hu
2021 arXiv   pre-print
Existing methods typically learn body and motion features from the body-joint trajectory, whereas they lack a systematic way to model body structure and underlying relations of body components beyond the  ...  from unlabeled skeleton graphs of various scales to learn an effective skeleton representation for person Re-ID.  ...  For the downstream task of person Re-ID, we extract encoded graph states ( ) learned from the pre-trained framework, and exploit an MLP ( (·)) to predict the sequence label.  ... 
arXiv:2107.01903v1 fatcat:5mduyl2xdjepxhyxlz43uo2ixe

3D Hand Pose Estimation via Graph-based Reasoning

Jae-Hun Song, Suk-Ju Kang
2021 IEEE Access  
We also present a hierarchical structure with six branches that independently estimate the position of the palm and five fingers by adding hand connections of each joint using graph reasoning based on  ...  Nevertheless, regressing joint coordinates is still a challenging task due to joint flexibility and self-occlusion.  ...  This method inspired from the multi-task mechanism.  ... 
doi:10.1109/access.2021.3061716 fatcat:qocemou2ivhg5brkgkzov5e7gy

Skeleton Motion Recognition Based on Multi-Scale Deep Spatio-Temporal Features

Kai Hu, Yiwu Ding, Junlan Jin, Liguo Weng, Min Xia
2022 Applied Sciences  
This paper proposes a novel multi-scale time sampling module and a deep spatiotemporal feature extraction module, which strengthens the receptive field of the feature map and strengthens the extraction  ...  In the task of human motion recognition, the overall action span is changeable, and there may be an inclusion relationship between action semantics.  ...  In the data-driven task, we can completely learn the graph from the target task, considering that the value in B k can be any value, which can represent not only the physical structure of the human body  ... 
doi:10.3390/app12031028 fatcat:qc3adyv62jemlewzygjwc5k6wa

Multi-Scale Adaptive Aggregate Graph Convolutional Network for Skeleton-Based Action Recognition

Zhiyun Zheng, Yizhou Wang, Xingjin Zhang, Junfeng Wang
2022 Applied Sciences  
In recent years, there is a trend of using graph convolutional networks (GCNs) to model the human skeleton into a spatio-temporal graph to explore the internal connections of human joints that has achieved  ...  learning.  ...  From the spatial dimension, an action often only has close connections with a few joints.  ... 
doi:10.3390/app12031402 fatcat:3ceyzh5tjbb3jbkvx4a5nbeytm

A Multi-Task Learning Approach for Human Activity Segmentation and Ergonomics Risk Assessment [article]

Behnoosh Parsa, Ashis G. Banerjee
2020 arXiv   pre-print
We propose a novel multi-task framework for HAE that utilizes a Graph Convolutional Network backbone to embed the interconnections between human joints in the features.  ...  We propose a new approach to Human Activity Evaluation (HAE) in long videos using graph-based multi-task modeling.  ...  For a graph with human skeletal structure, A is designed based on the anatomical connections among the joints. W ∈ R D×F is the weight matrix that is to be learned.  ... 
arXiv:2008.03014v2 fatcat:r5mb62aszvabfbjlztnvpwwuhi

Graph Sequential Network for Reasoning over Sequences [article]

Ming Tu, Jing Huang, Xiaodong He, Bowen Zhou
2020 arXiv   pre-print
In this paper, we consider a novel case where reasoning is needed over graphs built from sequences, i.e. graph nodes with sequence data.  ...  Recently Graph Neural Network (GNN) has been applied successfully to various NLP tasks that require reasoning, such as multi-hop machine reading comprehension.  ...  To build a graph on these sentences, we extracted named entities (NE) and noun phrases (NP) and the question, and two sentences are connected if 1) they come from the same document; 2) they come from the  ... 
arXiv:2004.02001v1 fatcat:7h5m6wnxynb3zp3mnc7jncgumm

Symbiotic Graph Neural Networks for 3D Skeleton-based Human Action Recognition and Motion Prediction [article]

Maosen Li, Siheng Chen, Xu Chen, Ya Zhang, Yanfeng Wang, Qi Tian
2019 arXiv   pre-print
For the backbone, we propose multi-branch multi-scale graph convolution networks to extract spatial and temporal features.  ...  The part-scale graphs integrate body-joints to form specific parts, representing high-level relations. Moreover, dual bone-based graphs and networks are proposed to learn complementary features.  ...  It employs parallel multi-scale GCN branches to treat high-order action differences for rich dynamics learning and also considers multi-scale graphs for spatial feature extraction.  ... 
arXiv:1910.02212v1 fatcat:suouaghedffsfaszejnp375yxy

Skeleton-Based Action Recognition using Multi-Scale and Multi-Stream Improved Graph Convolutional Network

Wang Li, Xu Liu, Zheng Liu, Feixiang Du, Qiang Zou
2020 IEEE Access  
The essential idea of the multi-scale approach is to extract the multi-scale features and learn their complementarity at each scale.  ...  Graph convolution networks (GCN) generalized from CNNs can extract features of non-Euclidean data [11] , [12] .  ... 
doi:10.1109/access.2020.3014445 fatcat:55gqhhvbsngevmxqbtoq7l7tmi

Learning Spatial Context with Graph Neural Network for Multi-Person Pose Grouping [article]

Jiahao Lin, Gim Hee Lee
2021 arXiv   pre-print
In this work, we formulate the grouping task as a graph partitioning problem, where we learn the affinity matrix with a Graph Neural Network (GNN).  ...  Current grouping approaches rely on learned embedding from only visual features that completely ignore the spatial configuration of human poses.  ...  To extract features for different joint types, the mapping function uses different set of parameters for each of the J joint types.  ... 
arXiv:2104.02385v1 fatcat:rjs57hpta5aozc3ps4iwgleslu

Skeleton Aware Multi-modal Sign Language Recognition [article]

Songyao Jiang, Bin Sun, Lichen Wang, Yue Bai, Kunpeng Li, Yun Fu
2021 arXiv   pre-print
Some efforts have been made to use hand detectors with pose estimators to extract hand key points and learn to recognize sign language via Neural Networks, but none of them outperforms RGB-based methods  ...  It is an essential yet challenging task since sign language is performed with the fast and complex movement of hand gestures, body posture, and even facial expressions.  ...  Without the graph reduction, the GCN model can hardly learn from the complex skeleton graph with too many nodes and edges.  ... 
arXiv:2103.08833v5 fatcat:dl7aebwtxbam5ftpjkgaaa5pae

Deep Structured Learning for Facial Action Unit Intensity Estimation

Robert Walecki, Ognjen Rudovic, Vladimir Pavlovic, Bjoern Schuller, Maja Pantic
2017 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
Modeling this structure is critical for improving the estimation performance; however, this performance is bounded by the quality of the input features extracted from face images.  ...  We show that joint learning of the deep features and the target output structure results in significant performance gains compared to existing deep structured models for analysis of facial expressions.  ...  Deep Facial Features: In our experiments, we first use a CNN to extract the feature map f d (x, W ) from an input image x, where the network parameters are defined by W .  ... 
doi:10.1109/cvpr.2017.605 dblp:conf/cvpr/WaleckiRPSP17 fatcat:e66unqqze5dghkaj76tbdifohe
« Previous Showing results 1 — 15 out of 35,361 results