Constraint-Aware Generative Re-ranking for Multi-Objective Optimization in Advertising Feeds

2026-03-04

AI Summary

This work addresses the challenge of re-ranking native advertisements in information feeds, which entails maximizing platform revenue while preserving user experience—a constrained combinatorial optimization problem. Existing generative approaches suffer from high inference latency and inadequate constraint handling. To overcome these limitations, we propose a constraint-aware generative re-ranking framework that uniquely integrates constraints directly into the neural decoding process. By combining bounded autoregressive generation, constraint-bound control, and reward pruning, our method unifies sequence generation with multi-objective optimization in an end-to-end trainable architecture. Evaluated on large-scale industrial datasets and online A/B tests, the proposed approach significantly improves both platform revenue and user engagement while adhering to stringent latency requirements.

Technology Category

• Natural Language Processing: Generation
• Constraint Satisfaction and Optimization: Constraint Learning and Acquisition
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications

Abstract

Optimizing reranking in advertising feeds is a constrained combinatorial problem, requiring simultaneous maximization of platform revenue and preservation of user experience. Recent generative ranking methods enable listwise optimization via autoregressive decoding, but their deployment is hindered by high inference latency and limited constraint handling. We propose a constraint-aware generative reranking framework that transforms constrained optimization into bounded neural decoding. Unlike prior approaches that separate generator and evaluator models, our framework unifies sequence generation and reward estimation into a single network. We further introduce constraint-aware reward pruning, integrating constraint satisfaction directly into decoding to efficiently generate optimal sequences. Experiments on large-scale industrial feeds and online A/B tests show that our method improves revenue and user engagement while meeting strict latency requirements, providing an efficient neural solution for constrained listwise optimization.

Problem

• constrained optimization
• reranking
• multi-objective optimization
• advertising feeds
• latency constraints

Innovation

• constraint-aware decoding
• generative reranking
• multi-objective optimization
• reward pruning
• listwise optimization