Video suggestion and discovery for youtube: taking random walks through the view graph

By Shumeet Baluja, Rohan Seth, D. Sivakumar, Yushi Jing, Jay Yagnik, Shankar Kumar, Deepak Ravichandran, Mohamed Aly. 2008

521 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Information Systems

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Information Systems

Citations in the corpus

Pixie: A System for Recommending 3+ Billion Items to 200+ Million Users in Real-Time (2017, background): “However, classical recommender systems that consider catalogs that only contain millions of items (movies [5, 18], videos [4, 7, 10], friends to follow [2, 14, 15]).”
Graph Embedding Techniques, Applications, and Performance: A Survey (2017, methodology): “Random walk based models [9], [10] propagate the labels with random walks.”
Related Pins at Pinterest: The Evolution of a Real-World Recommender System (2017, unknown): “Real-world recommender systems have been described for music suggestion [4], image search [12], video discovery on YouTube [2][6][7] and movies on Netﬂix [9].”
Down the (White) Rabbit Hole: The Extreme Right and Online Recommender Systems (2015, methodology): “For example, Baluja et al. (2008) suggested that standard approaches used in text domains were not easily applicable due to the difficulty of reliable video labeling. They proposed a graph-based approach that utilized the viewing patterns of YouTube users, which did not rely on the analysis of the u”
Social Networks and the Diffusion of User-Generated Content: Evidence from YouTube (2012, background): “For instance, the local popularity of a video is considerably different from that of global popularity (Baluja et al. 2008).”
Human Curation and Convnets: Powering Item-to-Item Recommendations on Pinterest (2015, methodology): “Collaborative ﬁltering [19] using user-generated signals (e.g. co-views, co-clicks) is widely used in commercially deployed recommendation systems such as YouTube related videos [1], [22], [2] and Amazon Related Items [16].”
Up next: retrieval methods for large scale related video suggestion (2014, methodology): “See Baluja et al. [7] (surveyed in Section 2.2) for the detailed description of an implementation of such a process.”
Social Network Data Analytics (2011, methodology): “Some recent advances in this direction are discussed in [1, 3].”
New Regularized Algorithms for Transductive Learning (2009, methodology): “Adsorption [1] is a general algorithmic framework for transductive learning where the learner is often given a small set of labeled examples and a very large set of unlabeled examples.”
Scaling Graph-based Semi Supervised Learning to Large Number of Labels Using Count-Min Sketch (2013, methodology): “Modified Adsorption (MAD) Modified Adsorption (MAD) [17] is a modification of an SSL method called Adsorption [2].”
Attentive Collaborative Filtering: Multimedia Recommendation with Item- and Component-Level Attention (2017, background): “Most of the current state-of-the-art multimedia recommendation techniques are based on the CF analysis [2, 4].”
A Distributed, Architecture-Centric Approach to Computing Accurate Recommendations from Very Large and Sparse Datasets (2011, background): “recommender systems do not attempt to replace, but rather complement the information retrieval research [8, 33, 76, 80, 120].”
Frame-Semantic Parsing (2014, background): “Traditionally, Markov random walks (Szummer and Jaakkola 2001; Baluja et al. 2008) or optimization of a loss function based on smoothness properties of the graph (e.g., Corduneanu and Jaakkola 2003; Zhu, Ghahramani, and Lafferty 2003; Subramanya and Bilmes 2008) are performed to propagate labels…”
Twitter Polarity Classification with Label Propagation over Lexical Links and the Follower Graph (2011, methodology): “Graph-based methods such as label propagation (Zhu and Ghahramani, 2002; Baluja et al., 2008; Talukdar and Crammer, 2009) provide a natural means to represent and exploit such relationships in order to improve classification, often while requiring less supervision than with standard classification.”