发表机构
DreamX, Alibaba Group（阿里巴巴集团 DreamX）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对多任务生成式推荐的三重崩溃问题，提出含DSD、TIM、HQ组件的IntHQ，在离线评估及高德地图生产环境中均取得推荐性能提升。
AI 中文摘要
异构数据上的多任务学习是现代推荐系统的基础，而生成模型正成为下一代推荐系统的核心。然而，将多任务学习融入生成范式的研究仍未得到充分探索。现有多任务推荐系统，无论是判别式还是生成式范式，都从单一的任务无关表示中提取任务相关特征，并将任务连接到预定义的转换流程中。我们表明，该方案本质上容易出现三重崩溃：源崩溃（任务特定信号注入较晚，在共享隐空间中被稀释）、关系崩溃（任务依赖要么被主干隐式吸收，要么被预定义流程静态固定）、分层崩溃（任务依赖不同尺度的特征，且在训练阶段发生变化）。我们提出IntHQ，这是一种多任务生成式推荐系统，包含三个组件，每个组件缓解一种崩溃：双流解耦（DSD）将任务标识尽早注入计算流，将共享上下文流与任务特定流分离，缓解信号稀释；任务交互建模（TIM）用显式跨任务交互取代预定义流程，让每个任务以学习到的、输入自适应的强度依赖于前序任务的实际输出；分层查询（HQ）让每个任务在不同训练阶段跨不同层收集多尺度信息。在离线评估中，IntHQ在四种代表性任务头配置下始终优于有竞争力的编码器主干；在高德地图（Amap）的生产环境中部署后，为亿级用户提供出行推荐服务，IntHQ带来了1.60%的相对UVCTR提升。
英文摘要
Multi-task learning over heterogeneous data is fundamental to modern recommendation, while generative models are emerging as the backbone of next-generation recommenders. However, the integration of multi-task learning into the generative paradigm remains largely unexplored. Existing multi-task recommenders, in both discriminative and generative paradigms, extract task-relevant features from a single task-agnostic representation and wire tasks into a predefined conversion funnel. We show that this scheme is inherently prone to a threefold collapse. Source collapse, where task-specific signals are injected late and diluted in the shared latent space. Relational collapse, where task dependencies are either implicitly absorbed by the backbone or statically fixed by predefined funnels. Hierarchical collapse, where tasks depend on features at different scales and shift across training stages. We propose IntHQ, a multi-task generative recommender with three components, each alleviating one collapse. Dual-Stream Decoupling (DSD) injects task identity into computation stream early and separates the shared context stream from the task-specific stream, alleviating signal dilution. Task-Interactive Modeling (TIM) replaces the predefined funnel with explicit cross-task interaction, letting each task condition on the realized outcomes of its predecessors with learned, input-adaptive strength. Hierarchical Querying (HQ) lets each task gather multi-scale information across different layers at different training stages. In offline evaluations, IntHQ consistently outperforms competitive encoder backbones under four representative task-head configurations. Deployed in production on Amap, serving hundreds of millions of users for travel recommendation, IntHQ yields a 1.60\% relative UVCTR lift.
