Atomic Intent Reasoning: Bringing LLM Semantics to Industrial Cross-Domain Recommendations

2026-06-08

AI Summary

This work addresses key challenges in cross-domain recommendation—namely, the substantial semantic gap, noisy and large-scale user behavior sequences, and high online inference latency of large language models (LLMs)—by proposing the AIR framework. AIR uniquely shifts LLM inference to the offline phase and introduces atomic intent representations combined with an efficient online retrieval-and-composition mechanism to dynamically construct user intents. This approach preserves strong semantic understanding while dramatically improving online efficiency. Evaluated on multiple public benchmarks, AIR achieves state-of-the-art performance and demonstrates significant real-world impact: in A/B tests on Kuaishou’s e-commerce platform, it increased GMV by 3.446%, substantially improved core metrics, and accelerated inference speed by approximately 400×, enabling the practical deployment of LLM-powered semantic capabilities in industrial-scale cross-domain recommender systems.

Abstract

Cross-domain recommendation is a core problem in content-to-e-commerce platforms. Its objective is to leverage user interactions with content to infer potential purchasing intent on the e-commerce side, thereby enhancing conversion rates and commercial value. However, in real industrial scenarios, cross-domain recommendation faces multiple challenges: significant semantic gaps exist between different domains, and user cross-domain behavior sequences are often massive in scale and rich in noise. Although large language models (LLMs) possess powerful semantic understanding and reasoning capabilities, their millisecond-level inference latency makes direct application in online recommendation systems difficult. To address these issues, this paper introduces AIR (Atomic Intent Reasoning), an LLM-driven cross-domain recommendation framework designed for industrial-grade deployment. By migrating LLM inference to the offline phase and dynamically constructing user intent representations through efficient retrieval and composition during online operations, it achieves approximately 400* inference acceleration while maintaining semantic consistency. Experimental results across multiple public datasets demonstrate that our method achieves state-of-the-art performance in cross-domain recommendation tasks. Furthermore, large-scale online A/B testing conducted in Kuaishou E-commerce's real-world business scenarios shows that our approach delivers stable and significant improvements across multiple core business metrics, including a +3.446% increase in GMV, fully validating its effectiveness and practical value in industrial-scale recommendation systems.

Problem

• cross-domain recommendation
• semantic gap
• user intent inference
• large language models
• industrial recommendation systems

Innovation

• Atomic Intent Reasoning
• Cross-domain Recommendation
• Large Language Models
• Industrial Deployment
• Efficient Inference