Search Results for « Playlist Embeddings »

In recent years, music storage and consumption has shifted massively to digital platforms, where large-scale libraries of songs are stored along with their metadata.As a byproduct of this transformation, music is increasingly being organized and accessed in the form of playlists.User-curated playlists have become massively available online, and the challenge of automatically generating playlists has gained popularity in the music information retrieval community.In this paper, we build on link prediction for graphs to propose a flexible music playlist generation method.We transform a playlist dataset into a weighted graph of songs and posit a Poisson model on the count of transitions between songs, where the rate is modulated by the euclidean distance between song embeddings.Our method yields prediction results superior to common deterministic baselines, suggesting that the learned embeddings can be used to derive a meaningful notion of song similarity.

This paper describes the approach of the D2KLab team to the RecSys Challenge 2018 that focuses on the task of playlist completion. We propose an ensemble strategy of different recurrent neural networks leveraging pre-trained embeddings representing tracks, artists, albums, and titles as inputs. We also use lyrics from which we extract semantic and stylistic features that we fed into the network for the creative track. The RNN learns a probabilistic model from the sequences of items in the playlist, which is then used to predict the most likely tracks to be added to the playlist. Concerning the playlists without tracks, we implemented a fall-back strategy called Title2Rec that generates recommendations using only the playlist title. We optimized the RNN, Title2Rec, and the ensemble approach on a validation set, tuning hyper-parameters such as the optimizer algorithm, the learning rate, and the generation strategy. This approach is effective in predicting tracks for a playlist and flexible to include diverse types of inputs, but it is also computationally demanding in the training phase.

Audio embeddings of musical similarity are often used for music recommendations and autoplay discovery. These embeddings are typically learned using co-listen data to train a deep neural network, to provide consistent tripletloss distances. Instead of directly using these co-listen–based embeddings, we explore making recommendations based on a second, smaller embedding space of human-intelligible musical attributes. To do this, we use the co-listen–based audio embeddings as inputs to small attribute classifiers, trained on a small hand-labeled dataset. These classifiers map from the original embedding space to a new interpretable attribute coordinate system that provides a more useful distance measure for downstream applications. The attributes and attribute embeddings allow us to provide a search interface and more intelligible recommendations for music curators. We examine the relative performance of these two embedding spaces (the co-listen–audio embedding and the attribute embedding) for the mathematical separation of thematic playlists. We also report on the usefulness of recommendations from the attribute-embedding space to human curators for automatically extending thematic playlists.

A continuação automática de playlists foi investigada como um problema de aprendizado de representação e recuperação vetorial. O objetivo foi verificar se embeddings comportamentais, aprendidos com Word2Vec sobre playlists ordenadas, seriam capazes de capturar relações latentes entre faixas e apoiar a continuação automática de playlists por meio de recuperação vetorial. Utilizou-se o Spotify Million Playlist Dataset como fonte de dados, aplicando-se um pipeline de preparação para estruturar, limpar e filtrar as playlists. As playlists foram tratadas como sequências ordenadas de faixas, e o Word2Vec foi empregado para aprender embeddings de músicas a partir de padrões de coocorrência. Posteriormente, o Annoy foi utilizado para recuperar faixas próximas no espaço vetorial. A avaliação adotou protocolos inspirados no RecSys Challenge 2018, com rankings de 500 recomendações para diferentes tamanhos de prefixo e duas estratégias de agregação: média simples e média ponderada por recência. Os resultados indicaram melhor desempenho em prefixos curtos, especialmente com 5 músicas observadas. Concluiu-se que os embeddings aprendidos capturaram relações úteis entre faixas, embora a agregação por média tenha apresentado perda de especificidade em contextos mais longos.

Digital music platforms have transformed the listening experience through curated playlists and transitions, yet many transition creating systems primarily serve professional DJs with many manual features to be set while overlooking amateur performers and everyday listeners. In this study, we introduce DJ-AI, a novel framework that bridges this gap by analyzing detailed musical features to optimize song sequences and create harmonic transitions between those songs. Our approach employs graph based optimization techniques to efficiently arrange playlists by mapping song relationships and determining the best transition paths. Additionally, we integrate MusicGen—a generative model for generating coherent musical continuations—and MERT audio embedding model, which capture nuanced musical attributes, to enhance the smoothness of transitions. Experimental evaluations reveal that DJ-AI outperforms traditional crossfade methods in generating smooth and coherent transitions. This framework paves the way for AI-driven adaptive mixing solutions, making seamless music transitions more accessible to a broader audience.

Where does the “power” of platformization reside? As is widely recognized, platforms are matchmakers which interface between different markets or “sides.” This article analyzes platform power dynamics through three of the most important markets that Spotify—the leading audio streaming platform—is embedded within: the music market; the advertising market; and the finance market. It does so through the lens of the playlist. Playlists can be seen as a central example of how platforms like Spotify employ curation or “curatorial power” to mediate markets in the attempt to advance their own interests. At the same time, playlists are an outcome of the conflicting pressures and tensions between these markets. As such, they provide a lens through which to view broader structural dynamics within the platform economy. As this case study of Spotify demonstrates, platform “power” is an always unstable and shifting outcome of the ongoing attempt to coordinate between various markets and actors.

In this paper, we propose a hybrid Neural Collaborative Filtering (NCF) model trained with a multi-objective function to achieve a music playlist generation system. The proposed approach focuses particularly on the cold-start problem (playlists with no seed tracks) and uses a text encoder employing a Recurrent Neural Network (RNN) to exploit textual information given by the playlist title. To accelerate the training, we first apply Weighted Regularized Matrix Factorization (WRMF) as the basic recommendation model to pre-learn latent factors of playlists and tracks. These factors then feed into the proposed multi-objective optimization that also involves embeddings of playlist titles. The experimental study indicates that the proposed approach can effectively suggest suitable music tracks for a given playlist title, compensating poor original recommendation results made on empty playlists by the WRMF model.