Infer As You Train: A Symmetric Paradigm of Masked Generative for Click-Through Rate Prediction

2025-11-18

AI Summary

Existing CTR prediction generative models employ the generative paradigm solely during training for representation learning, yet revert to discriminative inference—introducing train-inference asymmetry and hindering full exploitation of generative capabilities. To address this, we propose Symmetric Masked Generative CTR (SGCTR), the first model to unify iterative masked generation across both training and inference stages. SGCTR performs self-supervised reconstruction of input features to explicitly capture feature dependencies and dynamically optimizes user behavioral representations to suppress noise. Its core innovation lies in end-to-end integration of generative modeling—enabling dynamic redefinition of features throughout the entire lifecycle. Extensive experiments on multiple public benchmarks demonstrate that SGCTR consistently outperforms state-of-the-art methods by significant margins, validating both the effectiveness and generalizability of the symmetric generative paradigm for CTR prediction.

Technology Category

• Natural Language Processing: Generation
• Machine Learning: Deep Generative Models & Autoencoders
• Computer Vision: Generative Adversarial Networks (GANs) for Vision

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

Generative models are increasingly being explored in click-through rate (CTR) prediction field to overcome the limitations of the conventional discriminative paradigm, which rely on a simple binary classification objective. However, existing generative models typically confine the generative paradigm to the training phase, primarily for representation learning. During online inference, they revert to a standard discriminative paradigm, failing to leverage their powerful generative capabilities to further improve prediction accuracy. This fundamental asymmetry between the training and inference phases prevents the generative paradigm from realizing its full potential. To address this limitation, we propose the Symmetric Masked Generative Paradigm for CTR prediction (SGCTR), a novel framework that establishes symmetry between the training and inference phases. Specifically, after acquiring generative capabilities by learning feature dependencies during training, SGCTR applies the generative capabilities during online inference to iteratively redefine the features of input samples, which mitigates the impact of noisy features and enhances prediction accuracy. Extensive experiments validate the superiority of SGCTR, demonstrating that applying the generative paradigm symmetrically across both training and inference significantly unlocks its power in CTR prediction.

Problem

• Addresses asymmetry between training and inference in CTR prediction
• Leverages generative capabilities during online inference to refine features
• Mitigates noisy feature impact to enhance prediction accuracy

Innovation

• Symmetric masked generative paradigm for CTR prediction
• Iterative feature redefinition during online inference
• Mitigating noisy features to enhance prediction accuracy