LiSA: Linear Stochastic Attention for Long-Range Visual Modeling
cs.CV
Submitted: 2026-11-03
Updated: 2026-11-04
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Sources
- Hydra Attention: Efficient Attention with Many Heads
- EfficientViT: Multi-Scale Linear Attention for High-Resolution Dense Prediction
- PolaFormer: Polarity-aware Linear Attention for Vision Transformers
- Multi-hop Attention Graph Neural Network
- Learning Advanced Self-Attention for Linear Transformers in the Singular Value Domain
- mixup: Beyond Empirical Risk Minimization
Related papers
- TEA: Text Encoder Alignment for Robust Concept Erasure in Text-to-Image Models
- Do Egocentric Video-Language Models Capture Both Hand- and Object-Centric Cues?
- Aura: Consistent Multi-Subject Video Generation via VLM-Grounded Semantic Alignment
- PVRF: All-in-one Adverse Weather Removal via Prior-modulated and Velocity-constrained Rectified Flow
- DLWM: Diverse Latent World Models for Efficient Multimodal Reasoning
- Out of Sight, Still in Mind: Token Compression for Omni-LLMs