Filters








67,001 Hits in 7.4 sec

Semantic Product Search for Matching Structured Product Catalogs in E-Commerce [article]

Jason Ingyu Choi, Surya Kallumadi, Bhaskar Mitra, Eugene Agichtein, Faizan Javed
2020 arXiv   pre-print
Compared to web documents, product catalogs are more structured and sparse due to multi-instance fields that encode heterogeneous aspects of products (e.g. brand name and product dimensions).  ...  Lastly, we present a detailed error analysis to show which types of queries benefited the most by fielded representations and structured matching.  ...  To accomplish this, we propose a structured matching module (SMM) that leverages multiple fielded representations to learn a structured matching function. Our method has two advantages.  ... 
arXiv:2008.08180v1 fatcat:oyoowx4q45bihdpvc3swtszyuy

HYDRA

Siyuan Liu, Shuhui Wang, Feida Zhu, Jinbo Zhang, Ramayya Krishnan
2014 Proceedings of the 2014 ACM SIGMOD international conference on Management of data - SIGMOD '14  
matching; (II) constructing structural consistency graph to measure the high-order structure consistency on users' core social structures across different platforms; and (III) learning the mapping function  ...  We study the problem of large-scale social identity linkage across different social media platforms, which is of critical importance to business intelligence by gaining from social data a deeper understanding  ...  Yet, the weights of the attributes used in the matching can be learned from large training data by probabilistic modeling.  ... 
doi:10.1145/2588555.2588559 dblp:conf/sigmod/LiuWZZK14 fatcat:osatik6fcfhbxcjugwvtvzvslq

Ontology and Clustering Based Heterogeneous Data Sources Integration

Abrar Omar Alkhamisi
2020 International Journal of Advanced Trends in Computer Science and Engineering  
Secondly, it applies the kernel-based similarity learning to compute the similarity between the heterogeneous data sources.  ...  Hence, it is significant to integrate the heterogeneous data sources into unified information-rich systems with the help of ontologies.  ...  Learning method. • To group the semantically similar data sources using an Enhanced clustering method.  ... 
doi:10.30534/ijatcse/2020/79942020 fatcat:eqg4nhf3azfwjh4noal3dn6sly

Learning from Multiple Datasets with Heterogeneous and Partial Labels for Universal Lesion Detection in CT [article]

Ke Yan, Jinzheng Cai, Youjing Zheng, Adam P. Harrison, Dakai Jin, Youbao Tang, Yuxing Tang, Lingyun Huang, Jing Xiao, Le Lu
2021 arXiv   pre-print
In this work, we aim to develop a universal lesion detection algorithm to detect a variety of lesions. The problem of heterogeneous and partial labels is tackled.  ...  However, due to the annotation cost, datasets in medical imaging are often either partially-labeled or small.  ...  In this paper, we tackle the heterogeneous and partial label problem to aid large-scale multi-source deep learning in the lesion detection task.  ... 
arXiv:2009.02577v3 fatcat:pfw4h3yq5ndtnctj26ufuak3sa

Exploring Parameter Space with Structured Noise for Meta-Reinforcement Learning

Hui Xu, Chong Zhang, Jiaxing Wang, Deqiang Ouyang, Yu Zheng, Jie Shao
2020 Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence  
ESNPS utilizes meta-learning and directly uses meta-policy parameters, which contain prior knowledge, as structured noises to perturb the base model for effective exploration in new tasks.  ...  To this end, we propose a novel Exploration with Structured Noise in Parameter Space (ESNPS) approach.  ...  It uses the heterogeneous features of bags to learn a multi-view coordinated dictionary (representation space) and individual encoding vectors of bags for each view.  ... 
doi:10.24963/ijcai.2020/432 dblp:conf/ijcai/XingYWD020 fatcat:m3tyazmuyzh4ng5ndqwevw34ci

Structured Learning from Heterogeneous Behavior for Social Identity Linkage

Siyuan Liu, Shuhui Wang, Feida Zhu
2015 IEEE Transactions on Knowledge and Data Engineering  
Index Terms-Social identity linkage, structured Learning, heterogeneous behavior, multi-resolution temporal information matching ! Platform S'  ...  matching against high noise and information missing, and the behavior similarity are described by multi-dimensional similarity vector for each user pair; (II) we build structure consistency models to  ...  The pre-matched labeled data is generated by our rule-based filtering, a much more sophisticated set of measures than existing methods, including partial username overlapping [1] , [2] , user attribute  ... 
doi:10.1109/tkde.2015.2397434 fatcat:vh2kmavkgrhcneipro2d2ydfii

Generalized Unsupervised Manifold Alignment

Zhen Cui, Hong Chang, Shiguang Shan, Xilin Chen
2014 Neural Information Processing Systems  
Based on the assumption that datasets of the same theme usually have similar manifold structures, GUMA is formulated into an explicit integer optimization problem considering the structure matching and  ...  The main benefits of this model include: (1) simultaneous discovery and alignment of manifold structures; (2) fully unsupervised matching without any pre-specified correspondences; (3) efficient iterative  ...  Acknowledgments This work is partially supported by Natural Science Foundation of China under contracts Nos. 61272319, 61222211, 61202297, and 61390510.  ... 
dblp:conf/nips/CuiCSC14 fatcat:vskds52pjngkln7dzdzagjmt7i

Hetero-Labeled LDA: A Partially Supervised Topic Model with Heterogeneous Labels [chapter]

Dongyeop Kang, Youngja Park, Suresh N. Chari
2014 Lecture Notes in Computer Science  
We propose Hetero-Labeled LDA (hLLDA), a novel semi-supervised topic model, which can learn from multiple types of labels such as document labels and feature labels (i.e., heterogeneous labels), and also  ...  resolves both the label heterogeneity and label partialness problems in a unified generative process. hLLDA can leverage different forms of supervision and discover semantically coherent topics by exploiting  ...  In the future, we plan to incorporate additional type of label information such as partial or full taxonomy of topics [13] .  ... 
doi:10.1007/978-3-662-44848-9_41 fatcat:sv4tdozw6narpgdjzna7osdjvq

Exploring Heterogeneous Data Lake based on Unified Canonical Graphs

Qin Yuan, Ye Yuan, Zhenyu Wen, He Wang, Chen Chen, Guoren Wang
2022 Proceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval  
A matching entity based keyword search algorithm is presented to find answers across multiple data sources.  ...  A data lake is a repository for massive raw and heterogeneous data, which includes multiple data models with different data schemas and query interfaces.  ...  Adam optimizer [12] is used to optimize MatEnNet. We set that loss parameters 𝜆 = 1.5, 𝜂 = 0.1. The learning rate is 0.001. The model 𝑀 used in keyword matching is W2Vec [15] or GloVe [17] .  ... 
doi:10.1145/3477495.3531759 fatcat:m65v6focxjdshlkld6y5bcrxqi

Omni-Seg+: A Scale-aware Dynamic Network for Pathological Image Segmentation [article]

Ruining Deng, Quan Liu, Can Cui, Tianyuan Yao, Jun Long, Zuhayr Asad, R. Michael Womick, Zheyu Zhu, Agnes B. Fogo, Shilin Zhao, Haichun Yang, Yuankai Huo
2022 arXiv   pre-print
To handle this scaling issue, prior studies have typically trained multiple segmentation networks in order to match the optimal pixel resolution of heterogeneous tissue types.  ...  of pseudo-labels is introduced to model the inter-scale correlation of unannotated tissue types into a single end-to-end learning paradigm; and (3) superior scale-aware generalization is evidenced by  ...  Due to the issue of partial labeling, most approaches rely on an integration strategy to learn single segmentation from one network.  ... 
arXiv:2206.13632v1 fatcat:sdupx5lc3va6fosmlo3pztoiym

Modality-Dependent Cross-Modal Retrieval Based on Graph Regularization

Guanhua Wang, Hua Ji, Dexin Kong, Na Zhang
2020 Mobile Information Systems  
At the same time, the semantic information of class labels is used to reduce the semantic gaps between different modalities data and realize the interdependence and interoperability of heterogeneous data  ...  Secondly, utilizing the internal structure of original feature space constructs an adjacent graph with semantic information constraints which can make different labels of heterogeneous data closer to the  ...  Acknowledgments is work was partially supported by the National Natural Science Foundation of China (grant nos. 61772322, 61572298, 61702310, and 61873151).  ... 
doi:10.1155/2020/4164692 fatcat:2ku7xp5x65bkhggvu7x7skspja

Learning-Based Approaches for Matching Web Data Entities

Hanna Köpcke, Andreas Thor, Erhard Rahm
2010 IEEE Internet Computing  
Experimental Evaluation Let's now examine how effective learning-based match strategies can solve different match tasks on heterogeneous Web data entities in comparison to manually tuned strategies with  ...  Matching Web Data Sources Entities from Web data sources are particularly challenging to match because they are often highly heterogeneous with limited data quality regarding, for example, consistency  ... 
doi:10.1109/mic.2010.58 fatcat:upptlbexpvb4nocaq4sl74z4iq

Manifold alignment for heterogeneous single-cell multi-omics data integration using Pamona [article]

Kai Cao, Yiguang Hong, Lin Wan
2020 bioRxiv   pre-print
We formulate this task as a partial manifold alignment problem and develop a partial Gromov-Wasserstein optimal transport framework to solve it.  ...  Simulation studies and applications to four real data sets demonstrate that Pamona can accurately identify shared and dataset-specific cells, as well as faithfully recover and align cellular structures  ...  Label Transfer Accuracy, which has been widely used in the transfer learning community and was adopted by UnionCom [9] and SCOT [10] , is used to measure the ability to transfer labels of cells from  ... 
doi:10.1101/2020.11.03.366146 fatcat:uxg6sw4crbgl3jzyjkk7wsjot4

TOAST results for OAEI 2012

Arkadiusz Jachnik, Andrzej Szwabe, Pawel Misiorek, Przemyslaw Walkowiak
2012 International Semantic Web Conference  
Due to the flexibility of the integrated tensor-based representation of heterogeneous data, TOAST is able to learn the semantics equivalence relation on the basis of partial matches data included in a  ...  ., requiring no user intervention) ontology matching tool. TOAST is based on one of the first tensor-based approaches to Statistical Relational Learning.  ...  We use a 3rd-order tensor as a data structure that is suitable to represent data provided as a set of RDF triples [4] , [9] .  ... 
dblp:conf/semweb/JachnikSMW12 fatcat:md24scxjkvcoziuwsiphqesrvu

GPSP

Wenyu Du, Shuai Yu, Min Yang, Qiang Qu, Jia Zhu
2018 Companion of the The Web Conference 2018 on The Web Conference 2018 - WWW '18  
In this paper, we propose GPSP, a novel Graph Partition and Space Projection based approach, to learn the representation of a heterogeneous network that consists of multiple types of nodes and links.  ...  Finally, we concatenate the projective vectors from bipartite subnetworks with the ones learned from homogeneous subnetworks to form the final representation of the heterogeneous network.  ...  Following the strategy in [1] , we try to match this label set with the author embeddings, and get 103,024 successfully matched author embeddings with their labels.  ... 
doi:10.1145/3184558.3186928 dblp:conf/www/DuY0QZ18 fatcat:cz3zd6ovzrfvdapuvu5e6e3ksi
« Previous Showing results 1 — 15 out of 67,001 results