Meta-Graph Based Recommendation Fusion over Heterogeneous Information Networks

Zhao, Huan; Yao, Quanming; Li, Jianda; Song, Yangqiu; Lee, Dik Lun

doi:10.1145/3097983.3098063

Cited by 491 publications

(268 citation statements)

References 34 publications

Supporting

Mentioning

267

Contrasting

Unclassified

Order By: Relevance

“…Recently, increasing research attention has been paid to heterogeneous networks, highlighting multityped nodes and connections. Their modeling of rich semantics in terms of both node contents and typed links enables the integration of real-world data from various sources and facilitates wide applications [22,13,30,31,33].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Similarity Modeling on Heterogeneous Networks via Automatic Path Discovery

Yang

Liu

Fang

et al. 2019

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Heterogeneous networks are widely used to model real-world semi-structured data. The key challenge of learning over such networks is the modeling of node similarity under both network structures and contents. To deal with network structures, most existing works assume a given or enumerable set of meta-paths and then leverage them for the computation of meta-path-based proximities or network embeddings. However, expert knowledge for given meta-paths is not always available, and as the length of considered meta-paths increases, the number of possible paths grows exponentially, which makes the path searching process very costly. On the other hand, while there are often rich contents around network nodes, they have hardly been leveraged to further improve similarity modeling. In this work, to properly model node similarity in content-rich heterogeneous networks, we propose to automatically discover useful paths for pairs of nodes under both structural and content information. To this end, we combine continuous reinforcement learning and deep content embedding into a novel semi-supervised joint learning framework. Specifically, the supervised reinforcement learning component explores useful paths between a small set of example similar pairs of nodes, while the unsupervised deep embedding component captures node contents and enables inductive learning on the whole network. The two components are jointly trained in a closed loop to mutually enhance each other. Extensive experiments on three real-world heterogeneous networks demonstrate the supreme advantages of our algorithm.

show abstract

Section: Introductionmentioning

confidence: 99%

“…Consider our toy example in Figure 1, which is a simple heterogeneous network constructed with the Yelp data similarly as done in [33]. We only consider five node types: businesses (B), users (U), locations (L), categories (C) and stars (S).…”

Section: Introductionmentioning

confidence: 99%

Similarity Modeling on Heterogeneous Networks via Automatic Path Discovery

Yang

Liu

Fang

et al. 2019

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…Networks are widely used to model relational data such as web pages with hyperlinks and people with social connections. Recently, increasing research attention has been paid to the heterogeneous information network (HIN), due to its power of accommodating rich semantics in terms of multi-typed nodes (vertices) and links (edges), which enables the integration of real-world data from various sources and facilitates wide downstream applications [1], [2], [3], [4], [5], [6].…”

Section: Introductionmentioning

confidence: 99%

“…As for the second problem, most algorithms rely on supervised learning towards specific tasks to tune the weights on different meta-graphs [8], [3], [20], [21], [5], [10], [11]. Their performance heavily relies on labeled data.…”

Section: Introductionmentioning

confidence: 99%

Meta-Graph Based HIN Spectral Embedding: Methods, Analyses, and Insights

Yang

Feng

Pan

et al. 2018

2018 IEEE International Conference on Data Mining (ICDM)

View full text Add to dashboard Cite

Heterogeneous information network (HIN) has drawn significant research attention recently, due to its power of modeling multi-typed multi-relational data and facilitating various downstream applications. In this decade, many algorithms have been developed for HIN modeling, including traditional similarity measures and recent embedding techniques. Most algorithms on HIN leverage meta-graphs or meta-paths (special cases of meta-graphs) to capture various semantics. Given any arbitrary set of meta-graphs, existing algorithms either consider them as equally important or study their different importance through supervised learning. Their performance largely relies on prior knowledge and labeled data. While unsupervised embedding has shown to be a fundamental solution for various homogeneous network mining tasks, for HIN, it is a much harder problem due to such a presence of various meta-graphs.In this work, we propose to study the utility of different meta-graphs, as well as how to simultaneously leverage multiple meta-graphs for HIN embedding in an unsupervised manner. Motivated by prolific research on homogeneous networks, especially spectral graph theory, we firstly conduct a systematic empirical study on the spectrum and embedding quality of different meta-graphs on multiple HINs, which leads to an efficient method of meta-graph assessment. It also helps us to gain valuable insight into the higher-order organization of HINs and indicates a practical way of selecting useful embedding dimensions. Further, we explore the challenges of combining multiple meta-graphs to capture the multi-dimensional semantics in HIN through reasoning from mathematical geometry and arrive at an embedding compression method of autoencoder with 2,1-loss, which finds the most informative meta-graphs and embeddings in an end-to-end unsupervised manner. Finally, empirical analysis suggests a unified workflow to close the gap between our meta-graph assessment and combination methods. To the best of our knowledge, this is the first research effort to provide rich theoretical and empirical analyses on the utility of meta-graphs and their combinations, especially regarding HIN embedding. Extensive experimental comparisons with various state-of-the-art neural network based embedding methods on multiple real-world HINs demonstrate the effectiveness and efficiency of our framework in finding useful meta-graphs and generating high-quality HIN embeddings.

show abstract

“…In most existing work in mining HINs, relationships are defined by the selected meta-paths [28] in the network. Recently, another type of definition, meta-graph, is proposed to measure the similarity of nodes in HINs [39]. • Structural description for node pairs.…”

mentioning

confidence: 99%

On relationship formation in heterogeneous information networks: An inferring method based on multilabel learning

Chen

et al. 2019

Statistical Analysis

View full text Add to dashboard Cite

This paper studies how relationships form in heterogeneous information networks (HINs). The objective is not only to predict relationships in a given HIN more accurately but also to discover the interdependency between different type of relationships. A new relationship prediction method MULRP based on multilabel learning (MLL in brief) is proposed. In MULRP, the types of relationship between two nodes are represented by the meta-paths between nodes and each type of relationship is given a label. Under the framework of MLL, any potential relationships including the target relationship can be predicted. Moreover, the method can output the reasonable dependency scores between relationships. Thus, more viable paths will be provided to facilitate the formation of new relationships. The proposed method is evaluated on two real datasets: The DBLP Computer Science Bibliography(abbr. DBLP) network and Twitter network. The experimental results show that by using heterogeneous information in a supervised MLL setting, MULRP achieves better performance in comparison to several baseline binary classification methods and a state-of-art relationship prediction method. KEYWORDSheterogeneous information networks, meta-path, multilabel learning, relationship prediction INTRODUCTIONMany complex systems in real world can be formalized as networks, where nodes represent objects and links represent interactions between objects [15]. Most of these networks are heterogeneous, which contain various type of objects and relations. For example, in the online social network (OSN) Twitter, there are different types of nodes like users, locations and tweets, and different types of links like write/written, follow/followed, check-in/checked-in, etc. As a key subtask in link mining and social network analysis, link prediction aims to predict the formation of links in future based on the current or historical network [14]. It has wide application in bibliographic networks, biological networks, OSNs, recommendation systems and so on. Link prediction can be regarded as a simple binary classification problem: For any two unconnected objects, predict whether the link exists (with a positive label) or not (with a negative label). The prediction methods can be based on structural properties of the network [22] or the attributes of nodes.Many of the previous link prediction methods are designed for homogeneous information networks where all nodes or links are of the same type. These networks are usually the simplification of real interacting systems by ignoring its heterogeneity. For example, the co-authorship network only contains the author object and the co-author relationship. It is actually derived from a bibliographic network like DBLP,

show abstract

Meta-Graph Based Recommendation Fusion over Heterogeneous Information Networks

Cited by 491 publications

References 34 publications

Similarity Modeling on Heterogeneous Networks via Automatic Path Discovery

Similarity Modeling on Heterogeneous Networks via Automatic Path Discovery

Meta-Graph Based HIN Spectral Embedding: Methods, Analyses, and Insights

On relationship formation in heterogeneous information networks: An inferring method based on multilabel learning

Contact Info

Product

Resources

About