One Model, Two Markets: Bid-Aware Generative Recommendation
2026-03-23 • Information Retrieval
Information RetrievalArtificial IntelligenceComputer Science and Game TheoryMachine Learning
AI summaryⓘ
The authors propose GEM-Rec, a new recommendation system that not only suggests items based on how relevant they are but also considers ad revenue by including commercial bids. They add special control tokens to help decide when to show ads separately from what items to show, learning from past user interactions. Their method also uses a bidding system during recommendation generation that favors higher-paying ads without needing to retrain the model. Tests show that GEM-Rec helps platforms better balance useful recommendations with making money from ads.
Generative Recommender SystemsSequential RecommendationSemantic IDsControl TokensAd MonetizationBid-Aware DecodingAllocation MonotonicityInference ProcessInteraction LogsCommercial Retrieval
Authors
Yanchen Jiang, Zhe Feng, Christopher P. Mah, Aranyak Mehta, Di Wang
Abstract
Generative Recommender Systems using semantic ids, such as TIGER (Rajput et al., 2023), have emerged as a widely adopted competitive paradigm in sequential recommendation. However, existing architectures are designed solely for semantic retrieval and do not address concerns such as monetization via ad revenue and incorporation of bids for commercial retrieval. We propose GEM-Rec, a unified framework that integrates commercial relevance and monetization objectives directly into the generative sequence. We introduce control tokens to decouple the decision of whether to show an ad from which item to show. This allows the model to learn valid placement patterns directly from interaction logs, which inherently reflect past successful ad placements. Complementing this, we devise a Bid-Aware Decoding mechanism that handles real-time pricing, injecting bids directly into the inference process to steer the generation toward high-value items. We prove that this approach guarantees allocation monotonicity, ensuring that higher bids weakly increase an ad's likelihood of being shown without requiring model retraining. Experiments demonstrate that GEM-Rec allows platforms to dynamically optimize for semantic relevance and platform revenue.