Music Recommender Systems (MRS) have long relied on an information-retrieval framing, where progress is measured mainly through accuracy on retrieval-oriented subtasks. While effective, this reductionist paradigm struggles to address the deeper question of what makes a good recommendation, and attempts to broaden evaluation, through user studies or fairness analyses, have had limited impact. The emergence of Large Language Models (LLMs) disrupts this framework: LLMs are generative rather than ranking-based, making standard accuracy metrics questionable. They also introduce challenges such as hallucinations, knowledge cutoffs, non-determinism, and opaque training data, rendering traditional train/test protocols difficult to interpret. At the same time, LLMs create new opportunities, enabling natural-language interaction and even allowing models to act as evaluators. This work argues that the shift toward LLM-driven MRS requires rethinking evaluation. We first review how LLMs reshape user modeling, item modeling, and natural-language recommendation in music. We then examine evaluation practices from NLP, highlighting methodologies and open challenges relevant to MRS. Finally, we synthesize insights-focusing on how LLM prompting applies to MRS, to outline a structured set of success and risk dimensions. Our goal is to provide the MRS community with an updated, pedagogical, and cross-disciplinary perspective on evaluation.

This study presents a novel Multi-Modal Graph Neural Network (MM-GNN) framework for socially aware music recommendation, designed to enhance personalization and foster community-based engagement. The proposed model introduces a fusion-free deep mutual learning strategy that aligns modality-specific representations from lyrics, audio, and visual data while maintaining robustness against missing modalities. A heterogeneous graph structure is constructed to capture both user-song interactions and user-user social relationships, enabling the integration of individual preferences with social influence. Furthermore, emotion-aware embeddings derived from acoustic and textual signals contribute to emotionally aligned recommendations. Experimental evaluations on benchmark datasets demonstrate that MM-GNN significantly outperforms existing state-of-the-art methods across various performance metrics. Ablation studies further validate the critical impact of each model component, confirming the effectiveness of the framework in delivering accurate and socially contextualized music recommendations.

We present TalkPlayData 2, a synthetic dataset for multimodal conversational music recommendation generated by an agentic data pipeline. In the proposed pipeline, multiple large language model (LLM) agents are created under various roles with specialized prompts and access to different parts of information, and the chat data is acquired by logging the conversation between the Listener LLM and the Recsys LLM. To cover various conversation scenarios, for each conversation, the Listener LLM is conditioned on a finetuned conversation goal. Finally, all the LLMs are multimodal with audio and images, allowing a simulation of multimodal recommendation and conversation. In the LLM-as-a-judge and subjective evaluation experiments, TalkPlayData 2 achieved the proposed goal in various aspects related to training a generative recommendation model for music. TalkPlayData 2 and its generation code are released at https://talkpl.ai/talkplaydata2.

Natural language interfaces offer a compelling approach for music recommendation, enabling users to express complex preferences conversationally. While Large Language Models (LLMs) show promise in this direction, their scalability in recommender systems is limited by high costs and latency. Retrieval-based approaches using smaller language models mitigate these issues but often rely on single-modal item representations, overlook long-term user preferences, and require full model retraining, posing challenges for real-world deployment. In this paper, we present JAM (Just Ask for Music), a lightweight and intuitive framework for natural language music recommendation. JAM models user-query-item interactions as vector translations in a shared latent space, inspired by knowledge graph embedding methods like TransE. To capture the complexity of music and user intent, JAM aggregates multimodal item features via cross-attention and sparse mixture-of-experts. We also introduce JAMSessions, a new dataset of over 100k user-query-item triples with anonymized user/item embeddings, uniquely combining conversational queries and user long-term preferences. Our results show that JAM provides accurate recommendations, produces intuitive representations suitable for practical use cases, and can be easily integrated with existing music recommendation stacks.

Recommendation systems have become essential in modern music streaming platforms, shaping how users discover and engage with songs. One common approach in recommendation systems is collaborative filtering, which suggests content based on the preferences of users with similar listening patterns to the target user. However, this method is less effective on media where interactions are sparse. Music is one such medium, since the average user of a music streaming service will never listen to the vast majority of tracks. Due to this sparsity, there are several challenges that have to be addressed with other methods. This review examines the current state of research in addressing these challenges, with an emphasis on the role of content filtering in mitigating biases inherent in collaborative filtering approaches. We explore various methods of song classification for content filtering, including lyrical analysis using Large Language Models (LLMs) and audio signal processing techniques. Additionally, we discuss the potential conflicts between these different analysis methods and propose avenues for resolving such discrepancies.

Music recommender systems frequently utilize network-based models to capture relationships between music pieces, artists, and users. Although these relationships provide valuable insights for predictions, new music pieces or artists often face the cold-start problem due to insufficient initial information. To address this, one can extract content-based information directly from the music to enhance collaborative-filtering-based methods. While previous approaches have relied on hand-crafted audio features for this purpose, we explore the use of contrastively pretrained neural audio embedding models, which offer a richer and more nuanced representation of music. Our experiments demonstrate that neural embeddings, particularly those generated with the Contrastive Language-Audio Pretraining (CLAP) model, present a promising approach to enhancing music recommendation tasks within graph-based frameworks.

Automatic music recommendation has become an increasingly relevant problem in recent years, since a lot of music is now sold and consumed digitally. Most recommender systems rely on collaborative filtering. However, this approach suffers from the cold start problem: it fails when no usage data is available, so it is not effective for recommending new and unpopular songs. In this paper, we propose to use a latent factor model for recommendation, and predict the latent factors from music audio when they cannot be obtained from usage data. We compare a traditional approach using a bag-of-words representation of the audio signals with deep convolutional neural networks, and evaluate the predictions quantitatively and qualitatively on the Million Song Dataset. We show that using predicted latent factors produces sensible recommendations, despite the fact that there is a large semantic gap between the characteristics of a song that affect user preference and the corresponding audio signal. We also show that recent advances in deep learning translate very well to the music recommendation setting, with deep convolutional neural networks significantly outperforming the traditional approach.

Recommender systems serve a dual purpose for users: sifting out inappropriate or mismatched information while accurately identifying items that align with their preferences. Numerous recommendation algorithms are designed to provide users with a personalized array of information tailored to their preferences. Nevertheless, excessive personalization can confine users within a "filter bubble". Consequently, achieving the right balance between accuracy and diversity in recommendations is a pressing concern. To address this challenge, exemplified by music recommendation, we introduce the Diversified Weighted Hypergraph music Recommendation algorithm (DWHRec). In the DWHRec algorithm, the initial connections between users and listened tracks are represented by a weighted hypergraph. Simultaneously, associations between artists, albums and tags with tracks are also appended to the hypergraph. To explore users' latent preferences, a hypergraph-based random walk embedding method is applied to the constructed hypergraph. In our investigation, accuracy is gauged by the alignment between the user and the track, whereas the array of recommended track types measures diversity. We rigorously compared DWHRec against seven state-of-the-art recommendation algorithms using two real-world music datasets. The experimental results validate DWHRec as a solution that adeptly harmonizes accuracy and diversity, delivering a more enriched musical experience. Beyond music recommendation, DWHRec can be extended to cater to other scenarios with similar data structures.

Music streaming services heavily rely on their recommendation engines to continuously provide content to their consumers. Sequential recommendation consequently has seen considerable attention in current literature, where state of the art approaches focus on self-attentive models leveraging contextual information such as long and short-term user history and item features; however, most of these studies focus on long-form content domains (retail, movie, etc.) rather than short-form, such as music. Additionally, many do not explore incorporating negative session-level feedback during training. In this study, we investigate the use of transformer-based self-attentive architectures to learn implicit session-level information for sequential music recommendation. We additionally propose a contrastive learning task to incorporate negative feedback (e.g skipped tracks) to promote positive hits and penalize negative hits. This task is formulated as a simple loss term that can be incorporated into a variety of deep learning architectures for sequential recommendation. Our experiments show that this results in consistent performance gains over the baseline architectures ignoring negative user feedback.

Hosting about 50 million songs and 4 billion playlists, there is an enormous amount of data generated at Spotify every single day - upwards of 600 gigabytes of data (harvard.edu). Since the algorithms that Spotify uses in recommendation systems is proprietary and confidential, code for big data analytics and recommendation can only be speculated. However, it is widely theorized that Spotify uses two main strategies to target users' playlists and personalized mixes that are infamous for their retention - exploration and exploitation (kaggle.com). This paper aims to appropriate the filtering using the approach of deep learning for maximum user likeability. The architecture achieves 98.57% and 80% training and validation accuracy respectively.