Negative Feedback for Music Personalization

TL;DR

This paper demonstrates that incorporating real negative feedback, such as user skips, into music recommendation models improves accuracy and reduces training time, outperforming models trained with only positive feedback and random negatives.

Contribution

It introduces the use of real negative feedback in training music recommenders, showing benefits over traditional positive-only or randomly negative sampling methods.

Findings

Using explicit negative feedback reduces training time by ~60%.

Adding user skips increases user coverage and slightly improves accuracy.

Too many random negatives can introduce false negatives, limiting performance gains.

Abstract

Next-item recommender systems are often trained using only positive feedback with randomly-sampled negative feedback. We show the benefits of using real negative feedback both as inputs into the user sequence and also as negative targets for training a next-song recommender system for internet radio. In particular, using explicit negative samples during training helps reduce training time by ~60% while also improving test accuracy by ~6%; adding user skips as additional inputs also can considerably increase user coverage alongside slightly improving accuracy. We test the impact of using a large number of random negative samples to capture a 'harder' one and find that the test accuracy increases with more randomly-sampled negatives, but only to a point. Too many random negatives leads to false negatives that limits the lift, which is still lower than if using true negative feedback. We also find that the test accuracy is fairly robust with respect to the proportion of different feedback types, and compare the learned embeddings for different feedback types.