Xinyi Zhang

Xinyi Zhang (张新驿)

DLB 625G, David C. Lam Building

Shaw Campus, Hong Kong Baptist University

Hong Kong SAR

csxyzhang [at] comp.hkbu.edu.hk

I am a final-year PhD candidate in the Department of Computer Science at Hong Kong Baptist University (HKBU), advised by Prof. Jianliang Xu. I am currently also a visiting scholar at Nanyang Technological University (NTU), hosted by Dr. Qichen Wang. Before joining HKBU, I received my B.Eng. in Electronics and Information Engineering from Huazhong University of Science and Technology (HUST) in 2021.

Research. My research lies at the intersection of database systems, machine learning and security, with a focus on designing algorithms and index structures for efficient and secure query processing. I am particularly interested in:

Secure Query Processing — oblivious query processing for data federation and outsourced cloud computing; scalable secure query processing for vector databases based on Trusted Execution Environments.
AI for Databases (AI4DB) — robust updatable learned indexes that remain efficient under real-world, dynamic workloads.
Vector Databases — distributed and privacy-preserving systems for high-dimensional similarity search.
Query Optimization — distribution-aware algorithms and learned structure-guided execution plans for complex queries.

My work has been published at top-tier database venues including SIGMOD. I am actively looking for research collaborations and industry / post-doc opportunities in the above areas — please feel free to reach out.

news

Mar 01, 2026	Started as a Visiting Scholar at Nanyang Technological University (NTU), hosted by Dr. Qichen Wang. Working on learned indexes for query optimization.
Nov 24, 2025	Our paper HIRE: A Hybrid Learned Index for Robust and Efficient Performance under Mixed Workloads has been accepted to SIGMOD 2026! 🎉
Nov 15, 2023	Our paper FedKNN: Secure Federated k-Nearest Neighbor Search has been accepted to SIGMOD 2024! 🎉

selected publications

SIGMOD’26
HIRE: A Hybrid Learned Index for Robust and Efficient Performance under Mixed Workloads

Xinyi Zhang, Liang Liang, Anastasia Ailamaki, and Jianliang Xu

Proceedings of the ACM on Management of Data (SIGMOD), Feb 2026

Abs DOI Bib HTML PDF Slides

Indexes are critical for efficient data retrieval and updates in modern databases. Recent advances in machine learning have led to the development of learned indexes, which model the cumulative distribution function of data to predict search positions and accelerate query processing. While learned indexes substantially outperform traditional structures for point lookups, they often suffer from high tail latency, suboptimal range query performance, and inconsistent effectiveness across diverse workloads. To address these challenges, this paper proposes HIRE, a hybrid in-memory index structure designed to deliver efficient performance consistently. HIRE combines the structural and performance robustness of traditional indexes with the predictive power of model-based prediction to reduce search overhead while maintaining worst-case stability. Experimental results on multiple real-world datasets demonstrate that HIRE outperforms both state-of-the-art learned indexes and traditional structures in range-query throughput, tail latency, and overall stability, achieving up to 41.7x higher throughput under mixed workloads and reducing tail latency by up to 98%.
@article{SIGMOD26:hire, title = {HIRE: A Hybrid Learned Index for Robust and Efficient Performance under Mixed Workloads}, author = {Zhang, Xinyi and Liang, Liang and Ailamaki, Anastasia and Xu, Jianliang}, journal = {Proceedings of the ACM on Management of Data (SIGMOD)}, volume = {4}, number = {1}, articleno = {43}, numpages = {25}, month = feb, year = {2026}, doi = {10.1145/3786657}, }
SIGMOD’24
FedKNN: Secure Federated k-Nearest Neighbor Search

Xinyi Zhang, Qichen Wang, Cheng Xu, Yun Peng, and Jianliang Xu

Proceedings of the ACM on Management of Data (SIGMOD), Mar 2024

Abs DOI Bib HTML PDF Slides

Nearest neighbor search is a fundamental task in various domains, such as federated learning, data mining, information retrieval, and biomedicine. With the increasing need to utilize data from different organizations while respecting privacy regulations, private data federation has emerged as a promising solution. However, it is costly to directly apply existing approaches to federated k-nearest neighbor (kNN) search with difficult-to-compute distance functions, like graph or sequence similarity. To address this challenge, we propose FedKNN, a system that supports secure federated kNN search queries with a wide range of similarity measurements. Our system is equipped with a new Distribution-Aware kNN (DANN) algorithm to minimize unnecessary local computations while protecting data privacy. We further develop DANN*, a secure version of DANN that satisfies differential obliviousness. Extensive evaluations show that FedKNN outperforms state-of-the-art solutions, achieving up to 4.8x improvement on federated graph kNN search and up to 2.7x improvement on federated sequence kNN search.
@article{SIGMOD24:fedknn, title = {FedKNN: Secure Federated k-Nearest Neighbor Search}, author = {Zhang, Xinyi and Wang, Qichen and Xu, Cheng and Peng, Yun and Xu, Jianliang}, journal = {Proceedings of the ACM on Management of Data (SIGMOD)}, volume = {2}, number = {1}, articleno = {11}, numpages = {26}, month = mar, year = {2024}, doi = {10.1145/3639266}, }