Anqi Li (李安琪)

I am a Ph.D. student in Information and Communication Engineering at Shanghai Jiao Tong University (SJTU), advised by Prof. Guo Lu and Prof. Wenjun Zhang. I received my M.S. degree from Beijing Jiaotong University and my B.S. degree from Wuhan University of Technology.

My research focuses on LLM post-training and AIGC — aligning and enhancing large models after pre-training, and building generative models for visual and multimodal content.

News

Jun 10, 2026	Our work You Can Mask More For Extremely Low-Bitrate Image Compression was accepted to Pattern Recognition.
May 27, 2026	Our work UniNote: A Unified Embedding Model for Multimodal Representation and Ranking was accepted to the KDD 2026 Ads Track.
Feb 19, 2026	Our work Unified Spatiotemporal Token Compression for Video-LLMs at Ultra-Low Retention was accepted to CVPR 2026.
Jan 07, 2026	Our work Towards Trustworthy Multimodal Moderation via Policy-Aligned Reasoning was accepted to KDD 2026.
Sep 17, 2025	Our work FlowCut: Rethinking Redundancy via Information Flow for Efficient Vision-Language Models was accepted to NeurIPS 2025.
May 31, 2025	I received the Outstanding Graduate and Outstanding Master’s Thesis honors from Beijing Jiaotong University.
Jan 21, 2025	Our work Once-for-All: Controllable Generative Image Compression with Dynamic Granularity Adaptation was accepted to ICLR 2025.

Education

Ph.D. in Information and Communication Engineering

Shanghai Jiao Tong University · Advisors: Prof. Guo Lu and Prof. Wenjun Zhang

Sep 2025 – Present

M.S. in Information and Communication Engineering

Beijing Jiaotong University · Advisors: Prof. Huihui Bai and Prof. Yao Zhao

Sep 2022 – Jun 2025

B.S. in Software Engineering

Wuhan University of Technology

Sep 2018 – Jun 2022

Experience

Research Intern

Shanghai Artificial Intelligence Laboratory (Shanghai AI Lab) · Foundational Research Lab

Aug 2025 – Present

Research Intern

Xiaohongshu · Applied Algorithms & Content Understanding Group

Jul 2024 – Jul 2025

Publications

† Corresponding author.

KDD

Towards Trustworthy Multimodal Moderation via Policy-Aligned Reasoning and Hierarchical Labeling

Anqi Li, Wenwei Jin, Jintao Tong, Pengda Qin, Weijia Li, and Guo Lu^†

In Proceedings of the ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 2026

TL;DR: A policy-aligned reasoning framework with hierarchical labeling that makes multimodal content moderation transparent and trustworthy.

First Author arXiv Code

First-author publication
ICLR

Once-for-All: Controllable Generative Image Compression with Dynamic Granularity Adaptation

Anqi Li, Feng Li, Yuxi Liu, Runmin Cong, Yao Zhao, and Huihui Bai^†

In International Conference on Learning Representations, 2025

TL;DR: One generative codec that adapts compression granularity on the fly, covering a wide range of bitrates with a single model.

First Author arXiv Code

First-author publication
NeurIPS

FlowCut: Rethinking Redundancy via Information Flow for Efficient Vision-Language Models

Jintao Tong, Wenwei Jin, Pengda Qin, Anqi Li, Yixiong Zou, Yuhong Li, Yuhua Li, and Ruixuan Li^†

In Advances in Neural Information Processing Systems, 2025

TL;DR: Rethinks visual token redundancy through information flow and cuts tokens where information stops flowing, accelerating VLMs with minimal accuracy loss.

arXiv Code
CVPR

Unified Spatiotemporal Token Compression for Video-LLMs at Ultra-Low Retention

Junhao Du, Jialong Xue, Anqi Li, Jincheng Dai, and Guo Lu^†

In IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2026

TL;DR: Unifies spatiotemporal pruning and clustering so Video-LLMs stay accurate while keeping only a few percent of visual tokens.

arXiv
KDD

UniNote: A Unified Embedding Model for Multimodal Representation and Ranking

Jinghan Zhao, Wenwei Jin^†, Anqi Li, Jintao Tong, Luya Mo, Jiawei Li, Bin Li, and Yao Hu

In ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 2026

TL;DR: A unified multimodal embedding model with contrastive SFT and RL-based ranking refinement, deployed for industrial item-to-item retrieval at Xiaohongshu.

arXiv
PR

You Can Mask More For Extremely Low-Bitrate Image Compression

Anqi Li, Feng Li, Jiaxin Han, Huihui Bai, Runmin Cong, Chunjie Zhang, Meng Wang, Weisi Lin, and Yao Zhao^†

Pattern Recognition, 2026

TL;DR: Masks more of the image than prior methods and lets a learned model inpaint it back, pushing image compression to extremely low bitrates.

First Author arXiv Code

First-author publication

Services

Reviewer, ECCV 2026