影响力指数
论文质量、代表作、近期表现、广度与样本量置信度综合计算
58.68/100
前 5.1%
全站排名 #3,298
发表论文11 篇
平均评分
年均产出5.5 篇/年
Xiangyu Zeng
研究方向
Long-form Video · Video Understanding · Computer Vision · Salient object detection
16
VideoChat-Flash: Hierarchical Compression for Long-Context Video Modeling
ICLR 2026Poster
26
Balancing the Experts: Unlocking LoRA-MoE for GRPO via Mechanism-Aware Rewards
ICLR 2026Poster
三作18
TokenSculpt: Pruning with Min-Max Spatio-Temporal Duplication for Video Grounding
ICLR 2026Rejected
21
FreeRet: MLLMs as Training-Free Retrievers
ICLR 2026Rejected
二作24
RIVER: A Real-Time Interaction Benchmark for Video LLMs
ICLR 2026Poster
27
UniFlow: A Unified Pixel Flow Tokenizer for Visual Understanding and Generation
ICLR 2026Poster
三作28
Learning Goal-Oriented Language-Guided Navigation with Self-Improving Demonstrations at Scale
ICLR 2026Rejected
9
Benchmarking Visual Knowledge in Multimodal Large Language Models
ICLR 2026Withdrawn
23
StreamForest: Efficient Online Video Understanding with Persistent Event Memory
NeurIPS 2025Spotlight
一作19
VideoChat-R1.5: Visual Test-Time Scaling to Reinforce Multimodal Reasoning by Iterative Perception
NeurIPS 2025Poster
38
TimeSuite: Improving MLLMs for Long Video Understanding via Grounded Tuning
ICLR 2025Poster
一作