用统一用户故事序列提升流媒体发现效果
TubiFM: Unified Item, Carousel, and Search Ranking for Streaming Discovery
- 将用户跨场景行为转化为序列化输入，共享一个模型处理推荐、轮播和搜索
- 在线测试中搜索观看时长提升3.9%，轮播提升0.30%，延迟降低至200ms
- 适合需要简化推荐系统架构的平台，尤其关注全链路发现体验
个性化发现系统通常为项目排序、轮播排序和搜索分别训练独立模型，尽管这些任务共享同一用户旅程中的互补信号：观看行为影响轮播与项目排序，搜索查询揭示意图（即使未匹配到内容），观看历史有助于理解搜索是重看、续看还是新发现。我们提出用户故事，一种序列化表示，将用户的跨表面历史——包括属性、会话、带表面和轮播上下文的观看事件以及搜索事件——转换为单一标记序列。通过交错预训练语言标记与领域特定事件标记，用户故事使异构的推荐与搜索任务可表达为在共享语法上的提示式下一步标记预测。TubiFM 是此方法的一个实例：基于 Llama 3.2 1B 的模型，在用户故事上训练，并通过提示实现项目、轮播或搜索结果的排序，无需任务专用架构。离线评估中，该单一模型在项目、轮播和搜索排序上均优于专业基线。在线 A/B 测试中，TubiFM 显著提升搜索总观看时长（TVT）3.9%，轮播 TVT 提升 0.30%。项目排序对 TVT 影响统计中性（+0.14%），但与成熟生产堆栈相当；在所有三项任务中，TubiFM 在 L40S GPU 上运行，将 p99 排序延迟从 500ms 降至 200ms。结果表明，共享用户故事可在提升发现效果的同时简化排名系统。
原文摘要 · Abstract (English)
Personalized discovery systems often train separate models for item ranking, carousel ranking, and search, even though these tasks expose complementary signals from the same viewer journey: watches shape carousel and item ranking, search queries reveal intent even when they do not lead to a catalog match, and watch history helps interpret search as rewatching, continuation, or new discovery. We introduce the user story, a serialized representation that turns a user's cross-surface history - attributes, sessions, watch events with surface and carousel context, and search events - into a single token sequence. By interleaving pretrained language tokens with domain-specific event tokens, user stories let heterogeneous recommendation and search tasks be expressed as prompted next-token prediction over a shared grammar. TubiFM is one instantiation of this approach: a Llama 3.2 1B-based model trained on user stories and prompted to rank items, carousels, or search results without task-specific architectures. In offline evaluation, this single model outperforms specialist baselines across item, carousel, and search ranking. In online A/B tests, TubiFM significantly improves search total viewing time (TVT) by $+3.9\%$ and carousel TVT by $+0.30\%$. Item ranking is statistically neutral on TVT ($+0.14\%$), but matches a mature production stack; across all three tasks, TubiFM serves on L40S GPUs and reduces p99 ranking latency from 500ms to 200ms. These results show that shared user stories can improve discovery while simplifying ranking systems.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。