通过重写查询与融合排序，提升个性化对话搜索的鲁棒性与效率。
CFDA & CLIP at TREC iKAT 2025: Enhancing Personalized Conversational Search via Query Reformulation and Rank Fusion
- 采用Best-of-N与倒数排名融合策略优化检索结果
- 重排与融合显著提升系统在双任务中的稳定性
- 适合关注对话搜索效率与鲁棒性的研究者
2025年TREC交互式知识辅助任务（iKAT）包含交互式与离线提交两种任务。前者要求系统在实时约束下运行，鲁棒性与效率与准确性同等重要；后者则可在预定义数据集上对段落排序与响应生成进行受控评估。为此，我们探索了查询重写与检索融合作为核心策略。基于Best-of-N选择与倒数排名融合（RRF）构建流水线以应对不同任务。结果显示，重排与融合提升了系统的鲁棒性，同时揭示了两类任务中有效性和效率之间的权衡。
原文摘要 · Abstract (English)
The 2025 TREC Interactive Knowledge Assistance Track (iKAT) featured both interactive and offline submission tasks. The former requires systems to operate under real-time constraints, making robustness and efficiency as important as accuracy, while the latter enables controlled evaluation of passage ranking and response generation with pre-defined datasets. To address this, we explored query rewriting and retrieval fusion as core strategies. We built our pipelines around Best-of-$N$ selection and Reciprocal Rank Fusion (RRF) strategies to handle different submission tasks. Results show that reranking and fusion improve robustness while revealing trade-offs between effectiveness and efficiency across both tasks.
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。