Challenges and Solutions for Bandits in the Wild: Warm-Started Mixture Bandits for Cross-Cohort Slate Recommendation

2026-09-29

AI Summary

This study addresses the challenges of cold-start user preference learning and recommendation redundancy caused by the rapid depletion of limited item catalogs. To this end, we propose CohortMix-TS, an algorithm that introduces a novel cross-cohort transfer warm-start mechanism to accelerate new user modeling by leveraging historical cohort priors. Furthermore, it integrates Thompson sampling with an inventory-aware deduplication strategy to dynamically construct diverse session lists. Both simulation studies and real-world campus game experiments demonstrate that the proposed method significantly improves short-term recommendation quality, effectively reduces cumulative user regret, and prevents premature item exhaustion. Overall, this work establishes a new paradigm for sequential recommendation under finite catalog constraints.

Abstract

Many recommender services repeatedly encounter cold-start cohorts, where new users arrive with little or no interaction history. This creates two challenges: learning user preferences quickly from limited feedback and sustaining useful recommendations when each user has a finite catalog that can become repetitive or depleted over time. We propose CohortMix-TS, a warm-started mixture bandit that learns latent user groups from earlier cohorts and uses available metadata to construct group-informed priors for new users. Starting from these fixed priors, the model personalizes independently as feedback from each user becomes available. Session slates combine Thompson sampling with diversity and inventory-depletion controls. We evaluate CohortMix-TS through simulation, semi-synthetic experiments, and a 25-day randomized in-the-wild deployment with 713 registered participants in a Campus Games quiz application. Our evaluations show that cross-cohort transfer improves early recommendation quality and user-level regret, while inventory-aware slate construction helps prevent premature exhaustion of preferred items. In the field deployment, treatment users also showed a larger early-to-late change in correctness than users receiving random recommendations. Together, these results show how warm-start transfer and inventory-aware recommendations can support personalization for short-lived, repeatedly cold-starting cohorts.

Problem

• cold-start
• recommender systems
• bandits
• slate recommendation
• inventory depletion

Innovation

• Mixture Bandits
• Thompson Sampling
• Cold-Start
• Cross-Cohort Transfer
• Slate Recommendation