Dongfang Zhao, Ph.D.
Assistant Professor of Computer ScienceUW Tacoma School of Engineering and TechnologyOffice hours: Tuesday by appointment, CP 225, Tacoma
Data Science AffiliateUniversity of Washington eScience InstituteOffice hours: Friday by appointment, PAB C610, Seattle
Professional Services
- Editorial BoardIEEE Transactions on Parallel and Distributed Systems (TPDS)
- Senior Program CommitteeAAAI 2027, AAAI 2026
- Technical ReviewerVLDB'27, IPDPS'27, zbMATH, IEEE TDSC, IEEE TKDE, IEEE TKDD, IEEE TIFS
- Book ErrataIntroduction to Algebraic Geometry (by Brendan Hassett, Brown University)
- Open-Source ContributorMeta/Facebook Faiss (CUDA kernel); Microsoft DiskANN (C++ main branch)
- Software Author CertEx (exact vector search); OctEx (approximate vector search)
Selected Publications
- NeurIPS 2026MCGI: Manifold-Consistent Graph Indexing for Billion-Scale Disk-Resident Vector Search (acceptance rate 25%)
- NeurIPS 2026HPE: Hallucinated Positive Entanglement for Backdoor Attacks in Federated Self-Supervised Learning (acceptance rate 25%)
- KDD 2026RAE: A Neural Network Dimensionality Reduction Method for Nearest Neighbors Preservation in Vector Search (acceptance rate 18%)
- HPDC 2026SIVF: GPU-Resident IVF Index for Streaming Vector Analytics (acceptance rate 14%)
- ICDE 2026QPAD: Quantile-Preserving Approximate Dimension Reduction for Nearest Neighbors Preservation in High-Dimensional Vector Search
- CVPR 2026BadRSSD: Backdoor Attacks on Regularized Self-Supervised Diffusion Models (Findings)
- WWW 2026Reliable Non-Leveled Homomorphic Encryption for Web Services (acceptance rate 20%)
- AAAI 2026Order-Preserving Dimension Reduction for Multimodal Semantic Embedding (acceptance rate 17%)
- ICDM 2025ProHD: Projection-Based Hausdorff Distance Approximation (acceptance rate 13%)
- ECAI 2025IPBA: Imperceptible Perturbation Backdoor Attack in Federated Self-Supervised Learning (acceptance rate 23%)
- SIGMOD 2023Toward Efficient Homomorphic Encryption for Outsourced Databases through Parallel Caching
- DSN 2023 ECHAIN: Securing Electronic Device Provenance through Privacy-Preserving Consortium Blockchains
- SC 2021BAASH: Lightweight, Efficient, and Reliable Blockchain-as-a-Service for HPC Systems
- ICDE 2021SciChain: Blockchain-enabled Lightweight and Efficient Data Provenance for Reproducible Scientific Computing
- AAAI 2020HDK: Toward High-Performance Deep-Learning-Based Kirchhoff Analysis (Oral; acceptance rate 5%)
- SC 2019Swift Machine Learning Model Serving Scheduling: A Region Based Reinforcement Learning Approach
- VLDB 2017Comparative Evaluation of Big-Data Systems on Scientific Image Analytics Workloads
Acknowledgments
I am deeply grateful to the U.S. National Science Foundation and Anthropic for their support, without which my research would not be possible at the University of Washington.