Publications · Dongxie Wen

Publications

My work on matrix sketching, streaming algorithms, and efficient learning. You can also find my papers on Google Scholar.

Concept illustration of S3: online sampling selects a throughput-admissible subset for Newton-step model updates.

Certified Streaming Updates for Machine Learning Models under Limited Throughput

Dongxie Wen, Lu Yi, Fengtai Li, Shaoqiu Zhang, Xiao Zhang, Zhewei Wei

IEEE TPAMI 2026Early AccessLatest

Dyadic Block Sketching framework with geometrically growing blocks and per-block error management.

Revisiting Matrix Sketching in Linear Bandits: Achieving Sublinear Regret via Dyadic Block Sketching

Dongxie Wen, Hanyan Yin, Xiao Zhang, Peng Zhao, Lijun Zhang, Zhewei Wei

ICLR 2026

[ Paper ][ arXiv ]

Concept illustration of AeroSketch: randomized sketching and snapshots support persistent, sliding-window and distributed streams.

AeroSketch: Near-Optimal Time Matrix Sketch Framework for Persistent, Sliding Window, and Distributed Streams

Hanyan Yin, Dongxie Wen, Jiajun Li, Zhewei Wei, Xiao Zhang, Peng Zhao, Zhi-Hua Zhou

SIGMOD 2026

[ Paper ][ arXiv ]

FORKS framework linking incremental kernel approximation, feature mapping, and second-order online kernel learning.

Fast Second-Order Online Kernel Learning through Incremental Matrix Sketching and Decomposition

Dongxie Wen, Xiao Zhang*, Zhewei Wei*, Chenping Hou, Shuai Li, Weinan Zhang

IJCAI 2025

[ Paper ][ arXiv ][ Code ][ Blog ]

DS-FD framework: dumped snapshots maintain a sketch of the current sliding window.

Optimal Matrix Sketching over Sliding Windows

Best Paper Candidate Award

Hanyan Yin, Dongxie Wen, Jiajun Li, Zhewei Wei*, Xiao Zhang, Zengfeng Huang, Feifei Li

VLDB 2024

[ Paper ][ arXiv ][ Code ]