乔逸凡
Yifan Qiao

Department of Computer Science
University of California, Santa Barbara

yifanqiao@cs.ucsb.edu

Resume [Feb 2024]


About Me

Hello! I am a fifth-year PhD candidate at UC Santa Barbara, under the guidance of Professor Tao Yang. I have gained valuable experience through three consecutive summer internships at Apple in 2021, 2022, and 2023, under the supervision of Dr. Hua Ouyang.

Prior to joining UCSB, I earned my bachelor’s degree at Tsinghua University in China. I was supervised with Professor Zhiyuan Liu and Professor Chenyan Xiong in the THU NLP group. I also interned in the R-PAD Lab at CMU with Professor David Held in Summer 2018.

Research Interest

My primary research area centers around information retrieval systems, with a particular focus on building an efficient yet effective retrieval system with sparse and dense retrievers.

I am also interested in personalized on-device retrieval, where the user information may be extremely limited due to privacy requirements.

Publications

Efficient Sparse Retrieval

  1. Representation Sparsification with Hybrid Thresholding for Fast SPLADE-based Document Retrieval

    Yifan Qiao, Yingrui Yang, Shanxiu He, and Tao Yang. SIGIR 2023.

    A hybrid thresholding scheme which sparsifies the state-of-the-art sparse retrieval model SPLADE by 2-3x with minimal relevance trade-off. [checkpoints]

  2. Optimizing Guided Traversal for Fast Learned Sparse Retrieval

    Yifan Qiao, Yingrui Yang, Haixin Lin, and Tao Yang. WebConf 2023.

    A retrieval algorithm that uses traditional ranking feature BM25 to guide the retrieval on learned sparse scores. This two-level pruning control scheme with BM25 guided index traversal can achieve up-to 6.5x speedup for SPLADE sparse retrieval. [code]

  3. Dual Skipping Guidance for Document Retrieval with Learned Sparse Representations

    Yifan Qiao, Yingrui Yang, Haixin Lin, Tianbo Xiong, Xiyue Wang, and Tao Yang. Arxiv.

  4. Window Navigation with Adaptive Probing for Executing BlockMax WAND

    Jinjin Shao, Yifan Qiao, Shiyu Ji, and Tao Yang. SIGIR 2022.

Personalized Search System

  1. On-Device Query Auto-Completion for Email Search

    Yifan Qiao, Otto Godwin, and Hua Ouyang. To be appear in ECIR 2024.

    The prototype of the Apple Mail personalized on-device query completion experience.

Re-ranking

  1. Balanced Knowledge Distillation with Contrastive Learning for Document Re-ranking

    Yingrui Yang, Shanxiu He, Yifan Qiao, and Tao Yang. ICTIR 2023.

  2. Lightweight Composite Re-Ranking for Efficient Keyword Search with BERT

    Yingrui Yang, Yifan Qiao, Jinjin Shao, Xifeng Yan, and Tao Yang. WSDM 2022.

  3. Compact Token Representations with Contextual Quantization for Efficient Document Re-ranking

    Yingrui yang, Yifan Qiao, and Tao Yang. ACL 2022.

  4. Understanding the Behaviors of BERT in Ranking

    Yifan Qiao, Chenyan Xiong, Zhenghao Liu, and Zhiyuan Liu. Arxiv.

Privacy Preserving Retrieval

  1. Privacy-aware Document Retrieval with Two-level Inverted Indexing

    Yifan Qiao, Shiyu Ji, Changhai Wang, Jinjin Shao, and Tao Yang. Information Retrieval Journal, 2023.

  2. Index Obfuscation for Oblivious Document Retrieval in a Trusted Execution Environment

    Jinjin Shao, Shiyu Ji, Alvin Oliver Glova, Yifan Qiao, Tao Yang, and Tim Sherwood. CIKM 2020.

Services

Awards

Misc