A Contextual Random Walk Model for Automated Playlist Generation

Ueda, Seiji; Keyaki, Atsushi; Miyazaki, Jun

doi:10.1109/wi.2018.00-66

Cited by 2 publications

(3 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This current work generalizes this methodology by introducing multiple graphs that induce a hierarchy and a similarity metric that is not symmetric (directed graph) and leverages multiple hops between two objects in a sequence of objects. This current work is also closely related to the recent work of Ueda et al [9] where a single directed graph with nodes that represent different objects such as tracks, genres and artists is constructed. Their methodology uses a single biased random walk to generate recommendations, but this bias is an external parameter.…”

Section: Related Workmentioning

confidence: 79%

“…While there is a large number of approaches and algorithms for multimedia recommendation, they often fall into one of two broad categories: content-based [3], [15] and context-based [4], [5], [16]. Within the context-based approach there are various graph-based techniques where a graph of objects is constructed and then leveraged for recommendations, often using random walks [8], [7], [9].…”

Section: Related Workmentioning

confidence: 99%

“…While a effective similarity metric is important, the object recommendation requires a model. A recently proposed and promising approach are graph models, where objects are nodes in the graph and edges encode relationships among them [7], [8], [9]. Random walks on such graphs are often used to generate or recommend content.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A Contextual Hierarchical Graph Model for Generating Random Sequences of Objects with Application to Music Playlists

Nunes

Leite

Figueiredo

2019

2019 8th Brazilian Conference on Intelligent Systems (BRACIS)

View full text Add to dashboard Cite

Recommending the right content in large scale multimedia streaming services is an important and challenging problem that has received much attention in the past decade.A key ingredient for successful recommendations is an effective similarity metric between two objects, and models that leverage the current context to constrain the recommendations. This work proposes a model for random object generation that introduces two key novel elements: (i) a similarity metric based on the distance between objects in a given object sequence, that is also used to measure similarity between meta-data associated with the objects, such as artists and genres; (ii) a hierarchical graph model with different graphs each associated with a different meta-data. A biased random walk in each graph that are coupled and synchronized dictate the random generation of objects, leveraging the current context to constrain randomness. The proposed model is fully parameterized from sequences of objects, requiring no external parameters or tuning. The model is applied to a large music dataset with over 1 million playlists generating a hierarchy with three layers (genre, artist, track). Results indicate its superiority in generating actual full playlists against two baseline models.

show abstract

Section: Related Workmentioning

confidence: 79%