Calibrated Recommendations with Contextual Bandits
read the original abstract
Spotify's Home page features a variety of content types, including music, podcasts, and audiobooks. However, historical data is heavily skewed toward music, making it challenging to deliver a balanced and personalized content mix. Moreover, users' preference towards different content types may vary depending on the time of day, the day of week, or even the device they use. We propose a calibration method that leverages contextual bandits to dynamically learn each user's optimal content type distribution based on their context and preferences. Unlike traditional calibration methods that rely on historical averages, our approach boosts engagement by adapting to how users interests in different content types varies across contexts. Both offline and online results demonstrate improved precision and user engagement with the Spotify Home page, in particular with under-represented content types such as podcasts.
This paper has not been read by Pith yet.
Forward citations
Cited by 4 Pith papers
-
Catching a Moving Subspace: Low-Rank Bandits Beyond Stationarity
SPSC recovers drifting low-rank subspaces from scalar rewards under known noise variance, bounded coupling, and full probe support, then achieves dynamic regret scaling as r sqrt(T) plus lower-order terms instead of d...
-
Bandit Simulation for Average Reward Inference
BSI fits an environment simulator from bandit data and propagates parameter uncertainty to produce asymptotically valid confidence intervals for mean reward under arbitrary evaluation policies, including black-box ada...
-
Bandits for Efficient Experimentation: Adapting to Control Group, Preferences, and Context Drifts
Dri-MED achieves Õ(κ d² log T / Δ̃) regret and Õ(d) constraint violations for drifting contextual bandits with personalized preferences and baseline constraints under practitioner-friendly assumptions.
-
The Bandit's Blind Spot: The Critical Role of User State Representation in Recommender Systems
Variations in user state embeddings for CMAB recommenders can improve performance more than changing the bandit algorithm, with no embedding or aggregation strategy dominating across datasets.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.