Learning music similarity from relative user ratings

@article{Wolff2013LearningMS,
  title={Learning music similarity from relative user ratings},
  author={Daniel Wolff and Tillman Weyde},
  journal={Information Retrieval},
  year={2013},
  volume={17},
  pages={109-136}
}
Computational modelling of music similarity is an increasingly important part of personalisation and optimisation in music information retrieval and research in music perception and cognition. The use of relative similarity ratings is a new and promising approach to modelling similarity that avoids well known problems with absolute ratings. In this article, we use relative ratings from the MagnaTagATune dataset with new and existing variants of state-of-the-art algorithms and provide the first… 
Comparative Music Similarity Modelling Using Transfer Learning Across User Groups
TLDR
The proposed Relative Information-Theoretic Metric Learning (RITML) algorithm adapts a Mahalanobis distance using an iterative application of the ITML algorithm, thereby extending it to relative similarity data and is introduced, the first openly available relative similarity dataset with user attributes.
Probabilistic approaches for music similarity using restricted Boltzmann machines
TLDR
This paper takes advantage of RBM as a probabilistic neural network to assign a true hypothesis “ x is more similar to y than to z ” with a higher probability and shows that an energy-based optimisation approach with bipolar RBM can achieve better performance than other methods, including support vector machine and machine learning rank which are the state-of-the-art for this dataset.
Deep ranking: Triplet MatchNet for music metric learning
TLDR
A deep neural network named Triplet MatchNet is proposed to learn metrics directly from raw audio signals of triplets of music excerpts with human-annotated relative similarity in a supervised fashion and significantly outperforms three state-of-the-art music metric learning methods.
Feature Preprocessing with RBMs for Music Similarity Learning
TLDR
This study tests feature preprocessing with Restricted Boltzmann Machines in combination with established methods for learning distance measures and shows that this preprocessing improves the overall generalisation results of the trained models.
Learning Sparse Representation from Multiple-source Data for Relative Similarity in Music
TLDR
The use of sparse multi-modal restricted Boltzmann machine (mulRBM) in learning features from multiple sources for relative similarity and the results show that sparse features are better than original features, PCA features and features learned from mul-RBM without sparsity.
TRIPLET MATCHNET FOR MUSIC METRIC LEARNING
TLDR
A deep neural network named Triplet MatchNet is proposed to learn metrics directly from raw audio signals of triplets of music excerpts with human-annotated relative similarity in a supervised fashion and significantly outperforms three state-of-the-art music metric learning methods.
Feature Preprocessing with Restricted Boltzmann Machines for Music Similarity Learning
TLDR
This study tests feature preprocessing with Restricted Boltzmann Machines in combination with established methods for learning distance measures and shows that this preprocessing improves the overall generalisation results of the trained models.
A Music Classification Model based on Metric Learning and Feature Extraction from MP3 Audio Files
TLDR
This work proposes a novel music classification model based on metric learning and feature extraction from MP3 audio files to make possible learning a personalized metric for each customer.
Using PCA and K-Means to Predict Likeable Songs from Playlist Information
TLDR
A collaborative filtering approach using two different models (K- means and hierarchical clustering) is used to separate playlist data into clusters for comparison and when tested on a small sample of users, the system recommended songs that were considered likeable by the users 60% of the time, while still finding Songs that were generally diverse.
...
1
2
3
...

References

SHOWING 1-10 OF 64 REFERENCES
Adapting Metrics for Music Similarity Using Comparative Ratings
TLDR
This paper presents an approach to use machine learning techniques for analysing user data that specifies song similarity, and explores the potential for learning generalisable similarity measures with two stateof-the-art algorithms for learning metrics.
Combining Sources of Description for Approximating Music Similarity Ratings
TLDR
The results show that genre data allow more effective learning of a metric than simple audio features, but a combination of both feature sets clearly outperforms either individual set.
Computational Models of Music Similarity and their Application in Music Information Retrieval
TLDR
The combination of different approaches is optimized and the largest evaluation of music similarity measures published to date is presented, which confirms that genrebased evaluations are suitable to efficiently evaluate large parameter spaces.
Adapting similarity on the MagnaTagATune database: effects of model and feature choices
TLDR
It is found that the binary genre data shows little correlation with the similarity data, but combined with audio features it clearly improves generalisation, and the effectiveness of the MLR algorithm in generalising to unknown data is evaluated.
Similarity Based on Rating Data
TLDR
An algorithm to measure the similarity of two multimedia objects, such as songs or movies, using users’ preferences, and shows how this approach works by measuring its performance using an objective metric based on whether the same artist performed both songs.
A Systematic Comparison of Music Similarity Adaptation Approaches
TLDR
This paper provides a systematic comparison of algorithms for metric learning and higher-level facet distance weighting on the MagnaTagATune dataset and a crossvalidation variant taking into account clip availability is presented.
From Low-Level to High-Level: Comparative Study of Music Similarity Measures
TLDR
This work proposes two distance measures based on tempo-related aspects and a high-level semantic measure based on regression by support vector machines of different groups of musical dimensions such as genre and culture, moods and instruments, or rhythm and tempo.
Supporting Folk-Song Research by Automatic Metric Learning and Ranking
TLDR
It is shown how a weighted linear combination of different basic similarity measures can be automatically adapted to a specific retrieval task by learning this metric based on a special type of constraints.
Learning a Metric for Music Similarity
TLDR
Five different principled ways to embed songs into a Euclidean metric space are described, each of the six approaches rotate and scale the raw feature space with a linear transform and tune the parameters of these models using a song-classification task with content-based features.
Distance Metric Learning: A Comprehensive Survey
TLDR
A number of techniques that are central to distance metric learning are discussed, including convex programming, positive semi-definite programming, kernel learning, dimension reduction, K Nearest Neighbor, large margin classification, and graph-based approaches.
...
1
2
3
4
5
...