Random walks on the click graph

Craswell, Nick; Szummer, Martin

doi:10.1145/1277741.1277784

Cited by 373 publications

(387 citation statements)

References 11 publications

Supporting

Mentioning

380

Contrasting

Order By: Relevance

“…The high-level idea of this embedding method is to find the nodes that are most likely to be reached by a Markov random walk [25] inside the graph starting from the initial candidate nodes R. We use random walks to denoise the initial set R by finding nodes that are "close" to the majority of these initial vertices, while suppressing the effect of the outliers in R. In order to perform the random walk, for each graph edge linking i to j we define the onestep transition probability from i to j in terms of the weights w ij (we remind the reader that the weights w ij are the visual distances computed during the graph construction). Specifically, for each edge (i, j) we define the probability of transitioning from node i at time t to node j at time t + 1 to be…”

Section: Random Walk (Rw)mentioning

confidence: 99%

Measuring Image Distances via Embedding in a Semantic Manifold

Chen

Torresani

2012

Computer Vision – ECCV 2012

View full text Add to dashboard Cite

Abstract. In this work we introduce novel image metrics that can be used with distance-based classifiers or directly to decide whether two input images belong to the same class. While most prior image distances rely purely on comparisons of low-level features extracted from the inputs, our metrics use a large database of labeled photos as auxiliary data to draw semantic relationships between the two images, beyond those computable from simple visual features. In a preprocessing stage our approach derives a semantic image graph from the labeled dataset, where the nodes are the labeled images and the edges connect pictures with related labels. The graph can be viewed as modeling a semantic image manifold, and it enables the use of graph distances to approximate semantic distances. Thus, we reformulate the task of measuring the semantic distance between two unlabeled pictures as the problem of embedding the two input images in the semantic graph. We propose and evaluate several embedding schemes and graph distance metrics. Our results on Caltech101, Caltech256 and ImageNet show that our distances consistently match or outperform the state-of-the-art in this field.

show abstract

Section: Random Walk (Rw)mentioning

confidence: 99%

Measuring Image Distances via Embedding in a Semantic Manifold

Chen

Torresani

2012

Computer Vision – ECCV 2012

View full text Add to dashboard Cite

show abstract

“…Usage information in the form of click-through data has been exploited [1]. When a user enters a query, the system can exploit the behaviour of previous users that issued a similar query.…”

Section: Implicit Feedback In Multimedia Information Retrievalmentioning

confidence: 99%

“…In video retrieval, the interaction sequence is a reasonable way to track the user's information need. Craswell and Szummer [1] represent the clickthrough data of an image retrieval system as a graph, where queries and documents are the nodes and links are the clickthrough data. We adopt also a graphbased approach, as it facilitates the representation of interaction sequences.…”

Section: Implicit Feedback In Multimedia Information Retrievalmentioning

confidence: 99%

“…Craswell and Szummer [1] exploit the clickthrough data with a random walk algorithm. The random walk computation will end, in theory, with a higher probability on those nodes that previous users found (implicitly) relevant after issuing the query (forward walk approach) or on those documents that represent the information need of the query (backward walk approach).…”

Section: Wdg Based Recommendation Algorithmsmentioning

confidence: 99%

See 1 more Smart Citation

Use of Implicit Graph for Recommending Relevant Videos: A Simulated Evaluation

Vallet¹,

Hopfgartner²,

Jose³

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. In this paper, we propose a model for exploiting community based usage information for video retrieval. Implicit usage information from a pool of past users could be a valuable source to address the difficulties caused due to the semantic gap problem. We propose a graph-based implicit feedback model in which all the usage information can be represented. A number of recommendation algorithms were suggested and experimented. A simulated user evaluation is conducted on the TREC VID collection and the results are presented. Analyzing the results we found some common characteristics on the best performing algorithms, which could indicate the best way of exploiting this type of usage information.

show abstract

“…Both ideas were further combined in a unified framework which considered also the bi-directional walk over hyperlinks (Shakery and Zhai, 2006). Random walks on graphs containing queries and clicked links (or entire search trails) were recently utilized for web search result expansion (Craswell and Szummer, 2007;. Searching with graph-based methods for typed entity classes on the Web was explored recently in several publications (Cheng et al, 2007;Tsikrika et al, 2007).…”

Section: Related Work On Link-based Analysismentioning

confidence: 99%

Search for expertise : going beyond direct evidence

Serdyukov

View full text Add to dashboard Cite

Random walks on the click graph

Cited by 373 publications

References 11 publications

Measuring Image Distances via Embedding in a Semantic Manifold

Measuring Image Distances via Embedding in a Semantic Manifold

Use of Implicit Graph for Recommending Relevant Videos: A Simulated Evaluation

Search for expertise : going beyond direct evidence

Contact Info

Product

Resources

About