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

• URL: http://arxiv.org/abs/2508.14500v2
• Date: Wed, 27 Aug 2025 04:19:38 GMT
• Title: DGenCTR: Towards a Universal Generative Paradigm for Click-Through Rate Prediction via Discrete Diffusion
• Authors: Moyu Zhang, Yun Chen, Yujun Jin, Jinxin Hu, Yu Zhang,
• Abstract summary: We propose a two-stage Discrete Diffusion-Based Generative CTR training framework (DGenCTR)<n>This two-stage framework comprises a diffusion-based generative pre-training stage and a CTR-targeted supervised fine-tuning stage for CTR.
• Score: 6.189010741030871
• License: http://creativecommons.org/licenses/by/4.0/
• 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.

Related papers

• Learnable Chernoff Baselines for Inference-Time Alignment [64.81256817158851]
  We introduce Learnable Chernoff Baselines as a method for efficiently and approximately sampling from exponentially tilted kernels.<n>We establish total-variation guarantees to the ideal aligned model, and demonstrate in both continuous and discrete diffusion settings that LCB sampling closely matches ideal rejection sampling.
  arXiv  Detail & Related papers  (2026-02-08T00:09:40Z)
• From Feature Interaction to Feature Generation: A Generative Paradigm of CTR Prediction Models [81.43473418572567]
  Click-Through Rate (CTR) prediction is a core task in recommendation systems.<n>We propose a novel generative framework to address embedding dimensional collapse and information redundancy.<n>We show that SFG consistently mitigates embedding collapse and reduces information redundancy, while yielding substantial performance gains.
  arXiv  Detail & Related papers  (2025-12-16T03:17:18Z)
• Infer As You Train: A Symmetric Paradigm of Masked Generative for Click-Through Rate Prediction [9.542597285477683]
  Generative models are increasingly being explored in click-through rate (CTR) prediction field.<n>Existing generative models typically confine the generative paradigm to the training phase.<n>We propose the Symmetric Masked Generative Paradigm for CTR prediction (SGCTR)<n>SGCTR applies the generative capabilities during online inference to iteratively mitigate the features of input samples.
  arXiv  Detail & Related papers  (2025-11-18T12:07:56Z)
• MATT-CTR: Unleashing a Model-Agnostic Test-Time Paradigm for CTR Prediction with Confidence-Guided Inference Paths [9.542597285477683]
  We propose a Model-Agnostic Test-Time paradigm (MATT) to unlock the predictive potential of trained CTR models.<n>To quantify the confidence of feature combinations, we introduce a hierarchical probabilistic hashing method.<n>We generate instance-specific inference paths through iterative sampling and aggregate the prediction scores from multiple paths to conduct robust predictions.
  arXiv  Detail & Related papers  (2025-10-10T02:22:55Z)
• STEP: Structured Training and Evaluation Platform for benchmarking trajectory prediction models [7.927039780654076]
  We introduce STEP -- a new benchmarking framework that addresses limitations by providing a unified interface for multiple datasets.<n>We demonstrate the capabilities of STEP in a number of experiments which reveal 1) the limitations of widely-used testing procedures, 2) the importance of joint modeling of agents for better predictions of interactions, and 3) the vulnerability of current state-of-the-art models against both distribution shifts and targeted attacks by adversarial agents.
  arXiv  Detail & Related papers  (2025-09-18T09:56:16Z)
• Generative Click-through Rate Prediction with Applications to Search Advertising [6.555660969349762]
  We introduce a novel model that leverages generative models to enhance the precision of CTR predictions in discriminative models.<n>Our method's efficacy is substantiated through extensive experiments on a new dataset.<n>The model is deployed on one of the world's largest e-commerce platforms.
  arXiv  Detail & Related papers  (2025-07-15T12:21:30Z)
• Feature Staleness Aware Incremental Learning for CTR Prediction [18.810404607032297]
  Feature embeddings of a CTR prediction model often get stale when the corresponding features do not appear in current incremental data.<n>We propose a Feature Staleness Aware Incremental Learning method for CTR prediction (FeSAIL) which adaptively replays samples containing stale features.<n>FeSAIL outperforms various state-of-the-art methods on four benchmark datasets.
  arXiv  Detail & Related papers  (2025-04-29T09:05:47Z)
• Toward Guidance-Free AR Visual Generation via Condition Contrastive Alignment [31.402736873762418]
  Motivated by language model alignment methods, we propose textitCondition Contrastive Alignment (CCA) to facilitate guidance-free AR visual generation with high performance. Experimental results show that CCA can significantly enhance the guidance-free performance of all tested models with just one epoch fine-tuning. This experimentally confirms the strong theoretical connection between language-targeted alignment and visual-targeted guidance methods.
  arXiv  Detail & Related papers  (2024-10-12T03:31:25Z)
• Steering Masked Discrete Diffusion Models via Discrete Denoising Posterior Prediction [88.65168366064061]
  We introduce Discrete Denoising Posterior Prediction (DDPP), a novel framework that casts the task of steering pre-trained MDMs as a problem of probabilistic inference. Our framework leads to a family of three novel objectives that are all simulation-free, and thus scalable. We substantiate our designs via wet-lab validation, where we observe transient expression of reward-optimized protein sequences.
  arXiv  Detail & Related papers  (2024-10-10T17:18:30Z)
• Calibration of Time-Series Forecasting: Detecting and Adapting Context-Driven Distribution Shift [28.73747033245012]
  We introduce a universal calibration methodology for the detection and adaptation of context-driven distribution shifts. A novel CDS detector, termed the "residual-based CDS detector" or "Reconditionor", quantifies the model's vulnerability to CDS. A high Reconditionor score indicates a severe susceptibility, thereby necessitating model adaptation.
  arXiv  Detail & Related papers  (2023-10-23T11:58:01Z)
• RanPAC: Random Projections and Pre-trained Models for Continual Learning [59.07316955610658]
  Continual learning (CL) aims to learn different tasks (such as classification) in a non-stationary data stream without forgetting old ones. We propose a concise and effective approach for CL with pre-trained models.
  arXiv  Detail & Related papers  (2023-07-05T12:49:02Z)
• Enhancing Few-shot NER with Prompt Ordering based Data Augmentation [59.69108119752584]
  We propose a Prompt Ordering based Data Augmentation (PODA) method to improve the training of unified autoregressive generation frameworks. Experimental results on three public NER datasets and further analyses demonstrate the effectiveness of our approach.
  arXiv  Detail & Related papers  (2023-05-19T16:25:43Z)
• Distributionally Robust Models with Parametric Likelihood Ratios [123.05074253513935]
  Three simple ideas allow us to train models with DRO using a broader class of parametric likelihood ratios. We find that models trained with the resulting parametric adversaries are consistently more robust to subpopulation shifts when compared to other DRO approaches.
  arXiv  Detail & Related papers  (2022-04-13T12:43:12Z)
• Overcoming Mode Collapse with Adaptive Multi Adversarial Training [5.09817514580101]
  Generative Adversarial Networks (GANs) are a class of generative models used for various applications. GANs have been known to suffer from the mode collapse problem, in which some modes of the target distribution are ignored by the generator. We introduce a novel training procedure that adaptively spawns additional discriminators to remember previous modes of generation.
  arXiv  Detail & Related papers  (2021-12-29T05:57:55Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.