Quick Lists: Enriched Playlist Embeddings for Future Playlist Recommendation

TL;DR

This paper introduces a new method for creating enriched playlist embeddings that account for playlist length, track order, and user information, improving playlist recommendation accuracy and addressing cold start issues.

Contribution

The paper proposes a novel playlist embedding technique that captures ordering, sequencing, and user context, enhancing recommendation quality and cold start handling.

Findings

Embeddings improve next-best playlist recommendations.

Side information helps mitigate cold start problems.

Embeddings are invariant to playlist length and sensitive to track order.

Abstract

Recommending playlists to users in the context of a digital music service is a difficult task because a playlist is often more than the mere sum of its parts. We present a novel method for generating playlist embeddings that are invariant to playlist length and sensitive to local and global track ordering. The embeddings also capture information about playlist sequencing, and are enriched with side information about the playlist user. We show that these embeddings are useful for generating next-best playlist recommendations, and that side information can be used for the cold start problem.