Storm-Bench: Evaluating Online Video QA under Evolving and Incomplete Evidence
Preprint, 2026
Temporal & Multimodal Intelligence
Hi there, I'm Siru Zhong, a Ph.D. Candidate at The Hong Kong University of Science and Technology, Guangzhou, advised by Prof. Yuxuan Liang and Prof. Yang Yue. I am currently visiting Tsinghua University and the National Supercomputing Center in Shenzhen, hosted by Prof. Haohuan Fu. I am also fortunate to receive guidance from Prof. James T. Kwok, Prof. Jeffrey Xu Yu, and Prof. Yangqiu Song from HKUST.
My research focuses on Temporal and Multimodal Intelligence. I study how learning systems retain and retrieve relevant history, integrate evolving evidence, and support reliable understanding of sequential and multimodal inputs. Beyond academia, I have worked on temporal data infrastructure at Tencent, multimodal perception at XPeng Motors, foundation models at Huawei, and video memory at Beta Infinity.
I am honored to be a consecutive recipient of the HKUST (GZ) DSA Excellent Research Award, a Tutorial Speaker for MM4ST @ MM 2025 & ICME 2026 and MM4TS @ AAAI 2026, a Web Chair for WebST 2025 & 2026 and UrbComp 2025, and a reviewer/PC member for venues including NeurIPS, ICLR, KDD, ACM MM, AAAI, PAKDD, IEEE TPAMI, FARS, MILETS, IJITDM, and Neurocomputing.
",
which does not match the baseurl ("") configured in _config.yml.
baseurl in _config.yml to "".
Preprint, 2026
NeurIPS (Advances in Neural Information Processing Systems) 2025, Santiago, America Spotlight (688/21,575, 3.19%)
Preprint, 2026