Saved in:
Bibliographic Details
Main Authors: Chang, Yu-Cheng, Yeo, Guan-Wei, Eugene, Quah, Shih, Fan-Jie, Kuo, Yuan-Ching, Yu, Tsung-En, Hsu, Hung-Chun, Tsai, Ming-Feng, Wang, Chuan-Ju
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2509.15588
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909797515788288
author Chang, Yu-Cheng
Yeo, Guan-Wei
Eugene, Quah
Shih, Fan-Jie
Kuo, Yuan-Ching
Yu, Tsung-En
Hsu, Hung-Chun
Tsai, Ming-Feng
Wang, Chuan-Ju
author_facet Chang, Yu-Cheng
Yeo, Guan-Wei
Eugene, Quah
Shih, Fan-Jie
Kuo, Yuan-Ching
Yu, Tsung-En
Hsu, Hung-Chun
Tsai, Ming-Feng
Wang, Chuan-Ju
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2509_15588
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle CFDA & CLIP at TREC iKAT 2025: Enhancing Personalized Conversational Search via Query Reformulation and Rank Fusion
Chang, Yu-Cheng
Yeo, Guan-Wei
Eugene, Quah
Shih, Fan-Jie
Kuo, Yuan-Ching
Yu, Tsung-En
Hsu, Hung-Chun
Tsai, Ming-Feng
Wang, Chuan-Ju
Information Retrieval
Artificial Intelligence
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.
title CFDA & CLIP at TREC iKAT 2025: Enhancing Personalized Conversational Search via Query Reformulation and Rank Fusion
topic Information Retrieval
Artificial Intelligence
url https://arxiv.org/abs/2509.15588