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

TL;DR

DGenCTR introduces a novel two-stage discrete diffusion-based generative framework tailored for click-through rate prediction, effectively leveraging generative modeling to improve CTR estimation especially in label-scarce scenarios.

Contribution

It proposes a new sample-level generative paradigm for CTR prediction using discrete diffusion models, diverging from traditional sequence generation approaches.

Findings

Effective in improving CTR prediction accuracy

Validated through extensive offline and online experiments

Outperforms existing CTR models in various settings

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.