GRAB: An LLM-Inspired Sequence-First Click-Through Rate Prediction Modeling Paradigm

2026-02-02

AI Summary

Traditional deep learning-based recommendation models face limitations in performance, efficiency, generalization, and long-sequence modeling. This work proposes GRAB, an end-to-end generative CTR prediction framework inspired by large language models and grounded in a “sequence-first” paradigm for user behavior modeling. Its core innovation is the Causal Action-aware Multi-channel Attention (CamA) mechanism, which effectively captures temporal dynamics and action signals within user behavior sequences while enabling efficient scaling. Experimental results demonstrate that upon full-scale deployment, GRAB achieves a 3.49% increase in CTR and a 3.05% boost in revenue. Moreover, the model’s representational capacity scales nearly linearly with sequence length, highlighting its strong adaptability to extended user histories.

Technology Category

• Machine Learning: Causal Learning
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Natural Language Processing: Generation

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Traditional Deep Learning Recommendation Models (DLRMs) face increasing bottlenecks in performance and efficiency, often struggling with generalization and long-sequence modeling. Inspired by the scaling success of Large Language Models (LLMs), we propose Generative Ranking for Ads at Baidu (GRAB), an end-to-end generative framework for Click-Through Rate (CTR) prediction. GRAB integrates a novel Causal Action-aware Multi-channel Attention (CamA) mechanism to effectively capture temporal dynamics and specific action signals within user behavior sequences. Full-scale online deployment demonstrates that GRAB significantly outperforms established DLRMs, delivering a 3.05% increase in revenue and a 3.49% rise in CTR. Furthermore, the model demonstrates desirable scaling behavior: its expressive power shows a monotonic and approximately linear improvement as longer interaction sequences are utilized.

Problem

• Click-Through Rate Prediction
• Deep Learning Recommendation Models
• Long-Sequence Modeling
• Generalization
• User Behavior Sequences

Innovation

• Generative CTR Prediction
• Causal Action-aware Multi-channel Attention
• Sequence Modeling
• Large Language Model-inspired Recommendation
• Scalable Recommendation System