papers

Publications (20)

cs.CV2024

Agent Attention: On the Integration of Softmax and Linear Attention

Dongchen Han, Tianzhu Ye, Yizeng Han +5

The attention module is the key component in Transformers. While the global attention mechanism offers high expressiveness, its excessive computational cost restricts its applicabi…

cs.CV2023

Dynamic Perceiver for Efficient Visual Recognition

Yizeng Han, Dongchen Han, Zeyu Liu +7

Early exiting has become a promising approach to improving the inference efficiency of deep networks. By structuring models with multiple classifiers (exits), predictions for ``eas…

cs.CV2024

GSVA: Generalized Segmentation via Multimodal Large Language Models

Zhuofan Xia, Dongchen Han, Yizeng Han +3

Generalized Referring Expression Segmentation (GRES) extends the scope of classic RES to refer to multiple objects in one expression or identify the empty targets absent in the ima…

cs.CV2025

Linear Differential Vision Transformer: Learning Visual Contrasts via Pairwise Differentials

Yifan Pu, Jixuan Ying, Qixiu Li +7

Vision Transformers (ViTs) have become a universal backbone for both image recognition and image generation. Yet their Multi-Head Self-Attention (MHSA) layer still performs a quadr…

cs.CV2022

Contrastive Language-Image Pre-Training with Knowledge Graphs

Xuran Pan, Tianzhu Ye, Dongchen Han +2

Recent years have witnessed the fast development of large-scale pre-training frameworks that can extract multi-modal representations in a unified form and achieve promising perform…

cs.CV2022

Pseudo-Q: Generating Pseudo Language Queries for Visual Grounding

Haojun Jiang, Yuanze Lin, Dongchen Han +2

Visual grounding, i.e., localizing objects in images according to natural language queries, is an important topic in visual language understanding. The most effective approaches fo…

cs.CV2024

Efficient Diffusion Transformer with Step-wise Dynamic Attention Mediators

Yifan Pu, Zhuofan Xia, Jiayi Guo +9

This paper identifies significant redundancy in the query-key interactions within self-attention mechanisms of diffusion transformer models, particularly during the early stages of…

cs.CV2025

Step by Step Network

Dongchen Han, Tianzhu Ye, Zhuofan Xia +4

Scaling up network depth is a fundamental pursuit in neural architecture design, as theory suggests that deeper models offer exponentially greater capability. Benefiting from the r…

cs.CV2024

Bridging the Divide: Reconsidering Softmax and Linear Attention

Dongchen Han, Yifan Pu, Zhuofan Xia +6

Widely adopted in modern Vision Transformer designs, Softmax attention can effectively capture long-range visual information; however, it incurs excessive computational cost when d…

cs.CV2025

Vision Transformers are Circulant Attention Learners

Dongchen Han, Tianyu Li, Ziyi Wang +1

The self-attention mechanism has been a key factor in the advancement of vision Transformers. However, its quadratic complexity imposes a heavy computational burden in high-resolut…

cs.CV2026

Linear-Time Global Visual Modeling without Explicit Attention

Ruize He, Dongchen Han, Gao Huang

Existing research largely attributes the global sequence modeling capability of Transformers to the explicit computation of attention weights, a process that inherently incurs quad…

cs.LG2026

SiameseNorm: Breaking the Barrier to Reconciling Pre/Post-Norm

Tianyu Li, Dongchen Han, Zixuan Cao +7

The long-standing tension between Pre- and Post-Norm remains an open problem in Transformer architecture, reflecting a fundamental trade-off between training stability and represen…

cs.CV2024

Demystify Mamba in Vision: A Linear Attention Perspective

Dongchen Han, Ziyi Wang, Zhuofan Xia +7

Mamba is an effective state space model with linear computation complexity. It has recently shown impressive efficiency in dealing with high-resolution inputs across various vision…

cs.CV2026

Linearizing Vision Transformer with Test-Time Training

Yining Li, Dongchen Han, Zeyu Liu +3

While linear-complexity attention mechanisms offer a promising alternative to Softmax attention for overcoming the quadratic bottleneck, training such models from scratch remains p…

cs.CV2026

ViT: Unlocking Test-Time Training in Vision

Dongchen Han, Yining Li, Tianyu Li +6

Test-Time Training (TTT) has recently emerged as a promising direction for efficient sequence modeling. TTT reformulates attention operation as an online learning problem, construc…

cs.CV2023

FLatten Transformer: Vision Transformer using Focused Linear Attention

Dongchen Han, Xuran Pan, Yizeng Han +2

The quadratic computation complexity of self-attention has been a persistent challenge when applying Transformer models to vision tasks. Linear attention, on the other hand, offers…

cs.CV2026

LINA: Linear Autoregressive Image Generative Models with Continuous Tokens

Jiahao Wang, Ting Pan, Haoge Deng +4

Autoregressive models with continuous tokens form a promising paradigm for visual generation, especially for text-to-image (T2I) synthesis, but they suffer from high computational…

q-bio.MN2025

POINT: a web-based platform for pharmacological investigation enhanced by multi-omics networks and knowledge graphs

Zihao He, Liu Liu, Dongchen Han +10

Network pharmacology (NP) explores pharmacological mechanisms through biological networks. Multi-omics data enable multi-layer network construction under diverse conditions, requir…

cs.CV2024

VL-Trojan: Multimodal Instruction Backdoor Attacks against Autoregressive Visual Language Models

Jiawei Liang, Siyuan Liang, Man Luo +4

Autoregressive Visual Language Models (VLMs) showcase impressive few-shot learning capabilities in a multimodal context. Recently, multimodal instruction tuning has been proposed t…

cs.CV2023

OT-Attack: Enhancing Adversarial Transferability of Vision-Language Models via Optimal Transport Optimization

Dongchen Han, Xiaojun Jia, Yang Bai +3

Vision-language pre-training (VLP) models demonstrate impressive abilities in processing both images and text. However, they are vulnerable to multi-modal adversarial examples (AEs…