Semi-supervised Learning over Heterogeneous Information Networks by Ensemble of Meta-graph Guided Random Walks

@inproceedings{Jiang2017SemisupervisedLO,
  title={Semi-supervised Learning over Heterogeneous Information Networks by Ensemble of Meta-graph Guided Random Walks},
  author={He Jiang and Yangqiu Song and Chenguang Wang and Ming Zhang and Yizhou Sun},
  booktitle={IJCAI},
  year={2017}
}
Heterogeneous information network (HIN) is a general representation of many real world data. The difference between HIN and traditional homogeneous network is that the nodes and edges in HIN are with types. In many applications, we need to consider the types to make the decision more semantically meaningful. For annotationexpensive applications, a natural way is to consider semi-supervised learning over HIN. In this paper, we present a semi-supervised learning algorithm constrained by the types… 

Figures and Tables from this paper

Meta-Graph Based HIN Spectral Embedding: Methods, Analyses, and Insights
TLDR
This work proposes an embedding compression method of autoencoder with l2,1-loss, which finds the most informative meta-graphs and embeddings in an end-to-end unsupervised manner and suggests a unified workflow to close the gap between the authors' meta- graph assessment and combination methods.
Meta-GNN: Metagraph Neural Network for Semi-supervised learning in Attributed Heterogeneous Information Networks
TLDR
This paper proposes a novel metagraph convolution operation to extract features from local metagraph-structured neighborhoods, thus capturing semantic higher-order relationships in AHINs and proposes a proposed neural architecture Meta-GNN, which extracts features of diverse semantics by utilizing multiple metagraphs.
mg2vec: Learning Relationship-Preserving Heterogeneous Graph Representations via Metagraph Embedding
TLDR
This paper proposes the novel framework of mg2vec, which learns the embeddings for metagraphs and nodes jointly and significantly outperforms a suite of state-of-the-art baselines in relationship mining tasks including relationship prediction, search and visualization.
Leveraging Meta-path Contexts for Classification in Heterogeneous Information Networks
TLDR
This work proposes ConCH, a graph neural network model that formulates the classification problem as a multitask learning problem that combines semi-supervised learning with self-super supervised learning to learn from both labeled and unlabeled data.
Recommendation on Heterogeneous Information Network with Type-Sensitive Sampling
TLDR
This paper proposes a HIN based framework for recommendation tasks, where GNNs with a type-sensitive sampling to handle the heterogeneous and power-law graphs, and adopts schema-based attention to output the distribution of sampling over types.
Interpretable Relation Learning on Heterogeneous Graphs
TLDR
A novel interpretable relation learning model named IRL is proposed, which can not only predict whether relations exist between node pairs, but also make the inference more transparent and convincing.
Relation Structure-Aware Heterogeneous Graph Neural Network
TLDR
Experiments and comparisons, based on semi-supervised classification tasks on large scale heterogeneous networks with over a hundred types of edges, show that RSHN significantly outperforms state-of-the-arts.
User-Guided Clustering in Heterogeneous Information Networks via Motif-Based Comprehensive Transcription
TLDR
This work comprehensively transcribes the higher-order interaction signals to a series of tensors via motifs and proposes the MoCHIN model based on joint non-negative tensor factorization, which outperforms all baselines in three evaluation tasks under three different metrics.
Heterogeneous Network Representation Learning: Survey, Benchmark, Evaluation, and Beyond
TLDR
This work aims to provide a unified framework to deeply summarize and evaluate existing research on heterogeneous network embedding (HNE), and provides a generic paradigm for the systematic categorization and analysis over the merits of various existing HNE algorithms.
...
...

References

SHOWING 1-10 OF 49 REFERENCES
Relational retrieval using a combination of path-constrained random walks
TLDR
A novel learnable proximity measure is described which instead uses one weight per edge label sequence: proximity is defined by a weighted combination of simple “path experts”, each corresponding to following a particular sequence of labeled edges.
Transductive Classification on Heterogeneous Information Networks with Edge Betweenness-based Normalization
TLDR
A novel method for transductive classification on heterogeneous information networks composed of multiple types of vertices, based on the intuition that edges bridging across communities are less trustworthy, which is more effective than a state-of-the-art method, GNetMine.
Graph Regularized Transductive Classification on Heterogeneous Information Networks
TLDR
This paper considers the transductive classification problem on heterogeneous networked data which share a common topic and proposes a novel graph-based regularization framework, GNetMine, to model the link structure in information networks with arbitrary network schema and arbitrary number of object/link types.
Integrating meta-path selection with user-guided object clustering in heterogeneous information networks
TLDR
This work proposes to integrate meta-path selection with user-guided clustering to cluster objects in networks, where a user first provides a small set of object seeds for each cluster as guidance, and an effective and efficient iterative algorithm, PathSelClus, is proposed to learn the model.
Meta Structure: Computing Relevance in Large Heterogeneous Information Networks
TLDR
This paper proposes to use meta structure, which is a directed acyclic graph of object types with edge types connecting in between, to measure the proximity between objects, and develops three relevance measures based on meta structure.
Text Classification with Heterogeneous Information Network Kernels
TLDR
A novel text as network classification framework is presented, which introduces a structured and typed heterogeneous information networks (HINs) representation of texts, and a meta-path based approach to link texts that outperforms the state-of-the-art methods and other HIN-kernels.
KnowSim: A Document Similarity Measure on Structured Heterogeneous Information Networks
TLDR
This work proposes a method to represent a document as a typed heterogeneous information network (HIN), where the entities and relations are annotated with types, and proposes to use the meta-path defined in HIN to compute distance between documents.
PathSim: Meta Path-Based Top-K Similarity Search in Heterogeneous Information Networks
TLDR
Under the meta path framework, a novel similarity measure called PathSim is defined that is able to find peer objects in the network (e.g., find authors in the similar field and with similar reputation), which turns out to be more meaningful in many scenarios compared with random-walk based similarity measures.
Mining Heterogeneous Information Networks: Principles and Methodologies
  • Yizhou Sun, Jiawei Han
  • Computer Science
    Mining Heterogeneous Information Networks: Principles and Methodologies
  • 2012
TLDR
This semi-structured heterogeneous network modeling leads to a series of new principles and powerful methodologies for mining interconnected data, including: (1) rank-based clustering and classification; (2) meta-path-based similarity search and mining; (3) relation strength-aware mining, and many other potential developments.
Incorporating World Knowledge to Document Clustering via Heterogeneous Information Networks
TLDR
Experimental results on two text benchmark datasets show that incorporating world knowledge as indirect supervision can significantly outperform the state-of-the-art clustering algorithms as well as clustering algorithm enhanced with world knowledge features.
...
...