DRanker:用于Disney+主页的基于Transformer的多任务排序模型
DRanker: A Transformer-Based Multi-Task Ranking Model for the Disney+ Homepage
AI总结:
该研究提出基于Transformer的多任务排序模型DRanker,用于Disney+主页个性化推荐,经评估其简单设计更优,部署后提升核心互动指标与日活跃用户,并分享规模化排序系统的实用经验。
摘要:
我们提出了DRanker(Disney+ Ranker),这是一个基于Transformer的多任务排序模型,用于Disney+主页的个性化推荐。我们描述了基于丰富互动历史序列的仅编码器架构,搭配轻量级因果去偏机制,以及支持Transformer排序规模化部署的推理栈。通过迭代开发,我们系统评估了更复杂的替代方案,发现更简单的设计在离线指标、定性行为评估及生产环境中始终表现更优。DRanker已全球部署,在排序器驱动的所有场景中,使核心互动指标提升超1%,日活跃用户提升0.5%。我们分享了塑造规模化排序系统的数据工程、去偏及架构选择的实用经验。
英文摘要:
We present DRanker (Disney+ Ranker), a transformer-based multitask ranking model serving personalized recommendations on the Disney+ homepage. We describe the encoder-only architecture over enriched engagement-history sequences paired with a lightweight causal debiasing mechanism, and the inference stack that enables transformer-based ranking at scale. Through iterative development, we systematically evaluated more complex alternatives and found that simpler designs consistently outperformed them on offline metrics, qualitative behavioral evaluation, and in production. Deployed globally, DRanker drives over 1% lift in our key engagement metric and 0.5% lift in daily active users across all experiences powered by the ranker. We share practical lessons on data engineering, debiasing, and architectural choices that shaped our ranking system at scale.
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。