Unified tag analysis with multi-edge graph

Liu, Dong; Yan, Shuicheng; Rui, Yong; Zhang, Hong-Jiang

doi:10.1145/1873951.1873958

Cited by 46 publications

(23 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Development of new models to capture such heterogeneous data types and relations will greatly enhance the flexibility and representation power of the GSSL techniques. In addition, learning theories and algorithms that fully exploit the heterogeneous data and structures, beyond the initial interesting works in [5] and [33], will have significant impact in this field. Most of these works still consolidate diverse data and relations back to a single homogeneous type in the final inference stage.…”

Section: Graphs Of Heterogeneous Features and Relationsmentioning

confidence: 99%

Robust and Scalable Graph-Based Semisupervised Learning

2012

View full text Add to dashboard Cite

Graph-based semisupervised learning methods and new techniques for handling contaminated noisy labels, and gigantic data sizes for web applications, are reviewed in this paper.By Wei Liu, Jun Wang, Member IEEE, and Shih-Fu Chang, Fellow IEEE ABSTRACT | Graph-based semisupervised learning (GSSL) provides a promising paradigm for modeling the manifold structures that may exist in massive data sources in highdimensional spaces. It has been shown effective in propagating a limited amount of initial labels to a large amount of unlabeled data, matching the needs of many emerging applications such as image annotation and information retrieval. In this paper, we provide reviews of several classical GSSL methods and a few promising methods in handling challenging issues often encountered in web-scale applications. First, to successfully incorporate the contaminated noisy labels associated with web data, label diagnosis and tuning techniques applied to GSSL are surveyed. Second, to support scalability to the gigantic scale (millions or billions of samples), recent solutions based on anchor graphs are reviewed. To help researchers pursue new ideas in this area, we also summarize a few popular data sets and software tools publicly available. Important open issues are discussed at the end to stimulate future research.

show abstract

Section: Graphs Of Heterogeneous Features and Relationsmentioning

confidence: 99%

Robust and Scalable Graph-Based Semisupervised Learning

2012

View full text Add to dashboard Cite

show abstract

“…Nevertheless, using a single edge to model the relationship of two images is inadequate in practice, especially for the real-world images typically associated with multiple tags [16].To effectively search the visual content of the Internet, it is necessary to recognize the object categories present in an image [17]. In earlier works to object recognition are relying on strongly supervised learning of the visual appearance of object categories an.…”

Section: Related Workmentioning

confidence: 99%

Unsupervised Object Annotation through Context Analysis

Riad¹,

Elminir²,

Abdel-Ghany³

2013

IJAIS

View full text Add to dashboard Cite

The goal of object level annotation is to locate and identify instances of an object category within an image. Nowadays, Most of the current object level annotation systems annotate the object according to the visual appearance in the image. Recognizing an object in an image based visual appearance yield ambiguity in object detection due to appearance confusion for example "sky" object may be annotated as "water" according to similarity in visual appearance. As a result, these systems don't recognize the objects in an image accurately due to the lack of scene context. In the task of visual object recognition, scene context can play important role in resolving the ambiguities in object detection. In order to solve the ambiguity problem, this paper presents a new technique for a context based object level annotation that considers both the semantic context and spatial context analysis to reduce ambiguous in object annotation.

show abstract

“…However, the modeling is performed by connecting any two entities with a single kind of relation, which actually falls in the single edge graph and thus differs from our proposal here. In particular, the authors in [16] introduced a homogeneous multi-edge graph model to describe multiple parallel relations between the multi-label images, where an edge connects two similar local regions in different images. However, it can only handle the homogeneous entities (local regions) and relations (pairwise regional similarities), and hence cannot be applied into the heterogeneous scenarios.…”

Section: Related Workmentioning

confidence: 99%

Hybrid social media network

Liu

Chen

et al. 2012

Proceedings of the 20th ACM International Conference on Multimedia

Self Cite

View full text Add to dashboard Cite

Analysis and recommendation of multimedia information can be greatly improved if we know the interactions between the content, user, and concept, which can be easily observed from the social media networks. However, there are many heterogeneous entities and relations in such networks, making it difficult to fully represent and exploit the diverse array of information. In this paper, we develop a hybrid social media network, through which the heterogeneous entities and relations are seamlessly integrated and a joint inference procedure across the heterogeneous entities and relations can be developed. The network can be used to generate personalized information recommendation in response to specific targets of interests, e.g., personalized multimedia albums, target advertisement and friend/topic recommendation. In the proposed network, each node denotes an entity and the multiple edges between nodes characterize the diverse relations between the entities (e.g., friends, similar contents, related concepts, favorites, tags, etc). Given a query from a user indicating his/her information needs, a propagation over the hybrid social media network is employed to infer the utility scores of all the entities in the network while learning the edge selection function to activate only a sparse subset of relevant edges, such that the query information can be best propagated along the activated paths. Driven by the intuition that much redundancy exists among the diverse relations, we have developed a robust optimization framework based on several sparsity principles. We show significant performance gains of the proposed method over the state of the art in multimedia retrieval and recommendation using data crawled from social media sites. To the best of our knowledge, this is the first model supporting not only aggregation but also judicious selection of heterogeneous relations in the social media networks.

show abstract

Unified tag analysis with multi-edge graph

Cited by 46 publications

References 25 publications

Robust and Scalable Graph-Based Semisupervised Learning

Robust and Scalable Graph-Based Semisupervised Learning

Unsupervised Object Annotation through Context Analysis

Hybrid social media network

Contact Info

Product

Resources

About