Built with and by Teycir Ben Soltane•
How to Use•FAQ•GitHub•arXiv.org•
Share:
ArXivCSExplorer
☆☆Bookmarks🏆RSSHow to UseFAQ
Home/Authors/Xiao Liu

Xiao Liu

5 indexed papers

Recent (6 mo)
4
With code
0
Influential cites
0
Benchmarked
0

Publications per year

5
26

Top categories

NLP×4Vision×2Info Retrieval×2Crypto×1AI×1Multimedia×1Networking×1

Frequent co-authors

Xinxin Liu2×
Shiwei Gan2×
Yafeng Yin2×
Lei Xie2×
Sanglu Lu2×
Han Zhang2×

Research Timeline

2026
Mean Masked Autoencoder with Flow-Mixing for Encrypted Traffic Classification

The paper proposes Mean MAE (MMAE), a novel self-supervised pre-training framework that uses flow mixing and teacher-student distillation to improve encrypted traffic classification by capturing multi-granularity context.

Beyond Static Dialogues: Benchmarking Realistic, Heterogeneous, and Evolving Long-Term Memory

The paper introduces RHELM, a new benchmark designed to test LLMs' long-term memory by simulating realistic, complex, and evolving dialogues that integrate multiple heterogeneous data sources.

Beyond Static Dialogues: Benchmarking Realistic, Heterogeneous, and Evolving Long-Term Memory

The paper introduces RHELM, a new benchmark designed to test LLMs' long-term memory by simulating realistic, complex, and evolving dialogues that integrate multiple heterogeneous data sources.

InfoMerge: Information-aware Token Compression for Efficient Video Large Language Models

InfoMerge is a novel, training-free method that significantly compresses visual tokens for Video-LLMs by estimating temporal redundancy and allocating tokens based on content richness, achieving high efficiency with minimal performance loss.

InfoMerge: Information-aware Token Compression for Efficient Video Large Language Models

InfoMerge is a novel, training-free method that significantly compresses visual tokens for Video-LLMs by estimating temporal redundancy and allocating tokens based on content richness, achieving high efficiency with minimal performance loss.

Highlighted terms show continued research focus across papers

Papers

cs.CVcs.CLRecentJun 1, 2026

InfoMerge: Information-aware Token Compression for Efficient Video Large Language Models

Xinxin Liu, Shiwei Gan, Xiao Liu, Yafeng Yin +2 more

InfoMerge is a novel, training-free method that significantly compresses visual tokens for Video-LLMs by estimating temporal redundancy and allocating tokens based on content richness, achieving high…

View →
cs.CVcs.CLRecentJun 1, 2026

InfoMerge: Information-aware Token Compression for Efficient Video Large Language Models

Xinxin Liu, Shiwei Gan, Xiao Liu, Yafeng Yin +2 more

InfoMerge is a novel, training-free method that significantly compresses visual tokens for Video-LLMs by estimating temporal redundancy and allocating tokens based on content richness, achieving high…

View →
cs.CLcs.IRRecentMay 29, 2026

Beyond Static Dialogues: Benchmarking Realistic, Heterogeneous, and Evolving Long-Term Memory

Han Zhang, Zihao Tang, Xin Yu, Xiao Liu +7 more

The paper introduces RHELM, a new benchmark designed to test LLMs' long-term memory by simulating realistic, complex, and evolving dialogues that integrate multiple heterogeneous data sources.

View →
cs.CLcs.IRRecentMay 29, 2026

Beyond Static Dialogues: Benchmarking Realistic, Heterogeneous, and Evolving Long-Term Memory

Han Zhang, Zihao Tang, Xin Yu, Xiao Liu +7 more

The paper introduces RHELM, a new benchmark designed to test LLMs' long-term memory by simulating realistic, complex, and evolving dialogues that integrate multiple heterogeneous data sources.

View →
cs.CRcs.AIcs.MMMar 31, 2026

Mean Masked Autoencoder with Flow-Mixing for Encrypted Traffic Classification

Xiao Liu, Xiaowei Fu, Fuxiang Huang, Lei Zhang

The paper proposes Mean MAE (MMAE), a novel self-supervised pre-training framework that uses flow mixing and teacher-student distillation to improve encrypted traffic classification by capturing multi…

View →