MDL：通过分词实现大规模工业推荐系统的统一多分布学习
MDL: A Unified Multi-Distribution Learner in Large-scale Industrial Recommendation through Tokenization
AI总结：
MDL通过统一的多分布学习框架，利用分词技术提升大规模工业推荐系统中多场景和多任务学习的性能。
摘要：
工业推荐系统越来越多地采用多场景学习（MSL）和多任务学习（MTL）来处理多样化的用户交互和上下文，但现有方法存在两个关键缺点：（1）由于与复杂特征模块的交互有限，导致大规模模型参数利用率低下；（2）难以在统一框架中联合建模场景和任务信息。为了解决这些挑战，我们提出了一种统一的多分布学习（MDL）框架，受大语言模型（LLMs）中
英文摘要：
Industrial recommender systems increasingly adopt multi-scenario learning (MSL) and multi-task learning (MTL) to handle diverse user interactions and contexts, but existing approaches suffer from two critical drawbacks: (1) underutilization of large-scale model parameters due to limited interaction with complex feature modules, and (2) difficulty in jointly modeling scenario and task information in a unified framework. To address these challenges, we propose a unified \textbf{M}ulti-\textbf{D}istribution \textbf{L}earning (MDL) framework, inspired by the "prompting" paradigm in large language models (LLMs). MDL treats scenario and task information as specialized tokens rather than auxiliary inputs or gating signals. Specifically, we introduce a unified information tokenization module that transforms features, scenarios, and tasks into a unified tokenized format. To facilitate deep interaction, we design three synergistic mechanisms: (1) feature token self-attention for rich feature interactions, (2) domain-feature attention for scenario/task-adaptive feature activation, and (3) domain-fused aggregation for joint distribution prediction. By stacking these interactions, MDL enables scenario and task information to "prompt" and activate the model's vast parameter space in a bottom-up, layer-wise manner. Extensive experiments on real-world industrial datasets demonstrate that MDL significantly outperforms state-of-the-art MSL and MTL baselines. Online A/B testing on Douyin Search platform over one month yields +0.0626\% improvement in LT30 and -0.3267\% reduction in change query rate. MDL has been fully deployed in production, serving hundreds of millions of users daily.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
