Robust Semi-supervised Representation Learning for Graph-Structured Data

2019 ◽  
pp. 131-143 ◽  
Author(s):  
Lan-Zhe Guo ◽  
Tao Han ◽  
Yu-Feng Li
Keyword(s):  
Representation Learning ◽  
Structured Data

scholarly journals An adversarial learning approach for discovering social relations in human-centered information networks

2020 ◽  
Vol 2020 (1) ◽  
Author(s):  
Shicheng Cui ◽  
Qianmu Li ◽  
Shu-Ching Chen
Keyword(s):  
Social Relations ◽  
Structural Information ◽  
Representation Learning ◽  
Structured Data ◽  
Networked Systems ◽  
Generative Adversarial Networks ◽  
Textual Information ◽  
Domain Space ◽  
Adversarial Networks ◽  
Physical Spaces

Abstract The analytics on graph-structured data in cyber spaces has advanced many human-centered computing technologies. However, if only utilizing the structural properties, we might be prohibited from unraveling unknown social relations of nodes especially in the structureless networked systems. Up-to-date ways to unfold latent relationships from graph-structured data are network representation learning (NRL) techniques, but it is difficult for most existing ones to deal with the network-structureless situations due to the fact that they largely depend on the observed connections. With the ever-broader spectrum of human-centered networked systems, large quantities of textual information have been generated and collected from social and physical spaces, which may provide the clues of hidden social relations. In order to discover latent social relations from the accompanied text resources, this paper attempts to bridge the gap between text data and graph-structured data so that the textual information can be encoded to substitute for those incomplete structural information. Generative adversarial networks (GANs) are employed in the cross-modal framework to make the transformed data indistinguishable in graph-domain space and also capable of depicting structure-aware relationships with network homophily. Experiments conducted on three text-based network benchmarks demonstrate that our approach can reveal more realistic social relations from text-domain information compared against the state-of-the-art baselines.

scholarly journals Effective Decoding in Graph Auto-Encoder Using Triadic Closure

2020 ◽  
Vol 34 (01) ◽  
pp. 906-913
Author(s):  
Han Shi ◽  
Haozheng Fan ◽  
James T. Kwok
Keyword(s):  
Link Prediction ◽  
Representation Learning ◽  
Structured Data ◽  
Structured Prediction ◽  
The Other ◽  
Graph Structure ◽  
Convolutional Network ◽  
Node Clustering ◽  
Computationally Expensive ◽  
Triadic Closure

The (variational) graph auto-encoder and its variants have been popularly used for representation learning on graph-structured data. While the encoder is often a powerful graph convolutional network, the decoder reconstructs the graph structure by only considering two nodes at a time, thus ignoring possible interactions among edges. On the other hand, structured prediction, which considers the whole graph simultaneously, is computationally expensive. In this paper, we utilize the well-known triadic closure property which is exhibited in many real-world networks. We propose the triad decoder, which considers and predicts the three edges involved in a local triad together. The triad decoder can be readily used in any graph-based auto-encoder. In particular, we incorporate this to the (variational) graph auto-encoder. Experiments on link prediction, node clustering and graph generation show that the use of triads leads to more accurate prediction, clustering and better preservation of the graph characteristics.

scholarly journals Machine Learning on Graph-Structured Data

2021 ◽  
Author(s):  
Claudio D. T. Barros ◽  
Daniel N. R. da Silva ◽  
Fabio A. M. Porto
Keyword(s):  
Machine Learning ◽  
Biological Networks ◽  
Representation Learning ◽  
Structured Data ◽  
Graph Classification ◽  
Open Problems ◽  
Continuous Increase ◽  
Knowledge Graphs ◽  
Traditional Approaches

Several real-world complex systems have graph-structured data, including social networks, biological networks, and knowledge graphs. A continuous increase in the quantity and quality of these graphs demands learning models to unlock the potential of this data and execute tasks, including node classification, graph classification, and link prediction. This tutorial presents machine learning on graphs, focusing on how representation learning - from traditional approaches (e.g., matrix factorization and random walks) to deep neural architectures - fosters carrying out those tasks. We also introduce representation learning over dynamic and knowledge graphs. Lastly, we discuss open problems, such as scalability and distributed network embedding systems.

scholarly journals Improving Attention Mechanism in Graph Neural Networks via Cardinality Preservation

2020 ◽  
Author(s):  
Shuo Zhang ◽  
Lei Xie
Keyword(s):  
Neural Networks ◽  
Theoretical Analysis ◽  
Message Passing ◽  
Representation Learning ◽  
Attention Mechanism ◽  
Structured Data ◽  
Clear Understanding ◽  
Graph Classification ◽  
Competitive Performance ◽  
Graph Neural Networks

Graph Neural Networks (GNNs) are powerful for the representation learning of graph-structured data. Most of the GNNs use a message-passing scheme, where the embedding of a node is iteratively updated by aggregating the information from its neighbors. To achieve a better expressive capability of node influences, attention mechanism has grown to be popular to assign trainable weights to the nodes in aggregation. Though the attention-based GNNs have achieved remarkable results in various tasks, a clear understanding of their discriminative capacities is missing. In this work, we present a theoretical analysis of the representational properties of the GNN that adopts the attention mechanism as an aggregator. Our analysis determines all cases when those attention-based GNNs can always fail to distinguish certain distinct structures. Those cases appear due to the ignorance of cardinality information in attention-based aggregation. To improve the performance of attention-based GNNs, we propose cardinality preserved attention (CPA) models that can be applied to any kind of attention mechanisms. Our experiments on node and graph classification confirm our theoretical analysis and show the competitive performance of our CPA models. The code is available online: https://github.com/zetayue/CPA.

A Model for Structured Data Entry Based on Explicit Descriptional Knowledge

1994 ◽  
Vol 33 (05) ◽  
pp. 454-463 ◽  
Author(s):  
A. M. van Ginneken ◽  
J. van der Lei ◽  
J. H. van Bemmel ◽  
P. W. Moorman
Keyword(s):  
Domain Knowledge ◽  
Data Entry ◽  
Knowledge Bases ◽  
Structured Data ◽  
Research Quality ◽  
Free Text ◽  
Specific Knowledge ◽  
User Input ◽  
The One ◽  
Research Quality Assessment

Abstract:Clinical narratives in patient records are usually recorded in free text, limiting the use of this information for research, quality assessment, and decision support. This study focuses on the capture of clinical narratives in a structured format by supporting physicians with structured data entry (SDE). We analyzed and made explicit which requirements SDE should meet to be acceptable for the physician on the one hand, and generate unambiguous patient data on the other. Starting from these requirements, we found that in order to support SDE, the knowledge on which it is based needs to be made explicit: we refer to this knowledge as descriptional knowledge. We articulate the nature of this knowledge, and propose a model in which it can be formally represented. The model allows the construction of specific knowledge bases, each representing the knowledge needed to support SDE within a circumscribed domain. Data entry is made possible through a general entry program, of which the behavior is determined by a combination of user input and the content of the applicable domain knowledge base. We clarify how descriptional knowledge is represented, modeled, and used for data entry to achieve SDE, which meets the proposed requirements.

Structured Reporting of Medical Findings: Evaluation of a System in Gastroenterology

1992 ◽  
Vol 31 (04) ◽  
pp. 268-274 ◽  
Author(s):  
W. Gaus ◽  
J. G. Wechsler ◽  
P. Janowitz ◽  
J. Tudyka ◽  
W. Kratzer ◽  
...  
Keyword(s):  
Structured Data ◽  
Structured Reporting ◽  
Free Text ◽  
Clinical Documentation ◽  
Abdominal Sonography ◽  
Technical Examination ◽  
Medical Reports ◽  
Time Required ◽  
Upper Abdominal ◽  
Structured Approach

Abstract:A system using structured reporting of findings was developed for the preparation of medical reports and for clinical documentation purposes in upper abdominal sonography, and evaluated in the course of routine use. The evaluation focussed on the following parameters: completeness and correctness of the entered data, the proportion of free text, the validity and objectivity of the documentation, user acceptance, and time required. The completeness in the case of two clinically relevant parameters could be compared with an already existing database containing freely dictated reports. The results confirmed the hypothesis that, for the description of results of a technical examination, structured data reporting is a viable alternative to free-text dictation. For the application evaluated, there is even evidence of the superiority of a structured approach. The system can be put to use in related areas of application.

Structured Data Entry for Reliable Acquisition of Pharmacokinetic Data

1996 ◽  
Vol 35 (03) ◽  
pp. 261-264 ◽  
Author(s):  
T. Schromm ◽  
T. Frankewitsch ◽  
M. Giehl ◽  
F. Keller ◽  
D. Zellner
Keyword(s):  
Text Processing ◽  
Data Entry ◽  
Processing System ◽  
Database System ◽  
Structured Data ◽  
Pharmacokinetic Parameters ◽  
Pharmacokinetic Data ◽  
Probability Of Error ◽  
Error Frequency ◽  
Estimated Error

Abstract:A pharmacokinetic database was constructed that is as free of errors as possible. Pharmacokinetic parameters were derived from the literature using a text-processing system and a database system. A random data sample from each system was compared with the original literature. The estimated error frequencies using statistical methods differed significantly between the two systems. The estimated error frequency in the text-processing system was 7.2%, that in the database system 2.7%. Compared with the original values in the literature, the estimated probability of error for identical pharmacokinetic parameters recorded in both systems is 2.4% and is not significantly different from the error frequency in the database. Parallel data entry with a text-processing system and a database system is, therefore, not significantly better than structured data entry for reducing the error frequency.

Sign in / Sign up

Export Citation Format

Share Document