DGenCTR: Towards a Universal Generative Paradigm for Click-Through Rate Prediction via Discrete Diffusion

2025-08-20

AI Summary

Existing generative recommendation models primarily focus on sequential item generation and are ill-suited for click-through rate (CTR) prediction, which critically depends on fine-grained user-item interaction features. To alleviate the performance bottleneck of discriminative models under label-sparse scenarios, we propose DGenCTR—the first discrete diffusion-based generative framework tailored for CTR prediction. Our approach addresses this challenge through three key innovations: (1) introducing discrete diffusion into CTR pretraining to establish a sample-level generative paradigm; (2) explicitly modeling user-item cross features to tightly integrate generative capability with discriminative CTR estimation; and (3) adopting a two-stage training strategy—discrete diffusion pretraining followed by supervised fine-tuning. Extensive experiments on multiple public benchmarks and online A/B tests demonstrate that DGenCTR consistently achieves significant improvements in both CTR prediction accuracy and generalization performance.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Deep Generative Models & Autoencoders
• Natural Language Processing: Generation

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Recent advances in generative models have inspired the field of recommender systems to explore generative approaches, but most existing research focuses on sequence generation, a paradigm ill-suited for click-through rate (CTR) prediction. CTR models critically depend on a large number of cross-features between the target item and the user to estimate the probability of clicking on the item, and discarding these cross-features will significantly impair model performance. Therefore, to harness the ability of generative models to understand data distributions and thereby alleviate the constraints of traditional discriminative models in label-scarce space, diverging from the item-generation paradigm of sequence generation methods, we propose a novel sample-level generation paradigm specifically designed for the CTR task: a two-stage Discrete Diffusion-Based Generative CTR training framework (DGenCTR). This two-stage framework comprises a diffusion-based generative pre-training stage and a CTR-targeted supervised fine-tuning stage for CTR. Finally, extensive offline experiments and online A/B testing conclusively validate the effectiveness of our framework.

Problem

• Develops generative CTR prediction without discarding cross-features
• Proposes discrete diffusion model for sample-level generation paradigm
• Addresses label scarcity through generative pre-training and fine-tuning

Innovation

• Discrete diffusion generative model
• Two-stage training framework
• Sample-level generation paradigm