# Atlas: Multi-Scale Attention Improves Long Context Image Modeling

arXiv: https://arxiv.org/abs/2503.12355 · HTML: https://mlsanity.com/paper/2503.12355/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Builds on 26 of its 32 references that are in the map.
- First posted on arXiv 16 Mar 2025.

## Builds on (most cited first, 26 of 26)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://mlsanity.com/paper/2103.14030/) (2021)
- [Densely Connected Convolutional Networks](https://mlsanity.com/paper/1608.06993/) (2016)
- [Feature Pyramid Networks for Object Detection](https://mlsanity.com/paper/1612.03144/) (2016)
- [Mamba: Linear-Time Sequence Modeling with Selective State Spaces](https://mlsanity.com/paper/2312.00752/) (2023)
- [Training data-efficient image transformers & distillation through attention](https://mlsanity.com/paper/2012.12877/) (2020)
- [Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution](https://mlsanity.com/paper/2409.12191/) (2024)
- [Longformer: The Long-Document Transformer](https://mlsanity.com/paper/2004.05150/) (2020)
- [Pyramid Vision Transformer: A Versatile Backbone for Dense Prediction without Convolutions](https://mlsanity.com/paper/2102.12122/) (2021)
- [Expanding Performance Boundaries of Open-Source Multimodal Models with Model, Data, and Test-Time Scaling](https://mlsanity.com/paper/2412.05271/) (2024)
- [Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour](https://mlsanity.com/paper/1706.02677/) (2017)
- [Contrastive Multiview Coding](https://mlsanity.com/paper/1906.05849/) (2019)
- [Vision Mamba: Efficient Visual Representation Learning with Bidirectional State Space Model](https://mlsanity.com/paper/2401.09417/) (2024)
- [CrossViT: Cross-Attention Multi-Scale Vision Transformer for Image Classification](https://mlsanity.com/paper/2103.14899/) (2021)
- [Retentive Network: A Successor to Transformer for Large Language Models](https://mlsanity.com/paper/2307.08621/) (2023)
- [Twins: Revisiting the Design of Spatial Attention in Vision Transformers](https://mlsanity.com/paper/2104.13840/) (2021)
- [Early Convolutions Help Transformers See Better](https://mlsanity.com/paper/2106.14881/) (2021)
- [Focal Self-attention for Local-Global Interactions in Vision Transformers](https://mlsanity.com/paper/2107.00641/) (2021)
- [MambaVision: A Hybrid Mamba-Transformer Vision Backbone](https://mlsanity.com/paper/2407.08083/) (2024)
- [LongNet: Scaling Transformers to 1,000,000,000 Tokens](https://mlsanity.com/paper/2307.02486/) (2023)
- [Visformer: The Vision-friendly Transformer](https://mlsanity.com/paper/2104.12533/) (2021)
- [EdgeViTs: Competing Light-weight CNNs on Mobile Devices with Vision Transformers](https://mlsanity.com/paper/2205.03436/) (2022)
- [Long-Short Transformer: Efficient Transformers for Language and Vision](https://mlsanity.com/paper/2107.02192/) (2021)
- [FasterViT: Fast Vision Transformers with Hierarchical Attention](https://mlsanity.com/paper/2306.06189/) (2023)
- [When an Image is Worth 1,024 x 1,024 Words: A Case Study in Computational Pathology](https://mlsanity.com/paper/2312.03558/) (2023)

## Built on by (most cited first, 4 of 4)

- [Pillar-0: A New Frontier for Radiology Foundation Models](https://mlsanity.com/paper/2511.17803/) (2025)
- [Jolia: Concept-Level Vision-Language Alignment for 3D CT Contrastive Learning](https://mlsanity.com/paper/2606.24570/) (2026)
- [Enhancing Fine-Grained Spatial Grounding in 3D CT Report Generation via Discriminative Guidance](https://mlsanity.com/paper/2604.10437/) (2026)
- [UCSF-PDGM-VQA: Visual Question Answering dataset for brain tumor MRI interpretation](https://mlsanity.com/paper/2605.17140/) (2026)

## Abstract (arXiv)

Efficiently modeling massive images is a long-standing challenge in machine learning. To this end, we introduce Multi-Scale Attention (MSA). MSA relies on two key ideas, (i) multi-scale representations (ii) bi-directional cross-scale communication. MSA creates O(log N) scales to represent the image across progressively coarser features and leverages cross-attention to propagate information across scales. We then introduce Atlas, a novel neural network architecture based on MSA. We demonstrate that Atlas significantly improves the compute-performance tradeoff of long-context image modeling in a high-resolution variant of ImageNet 100. At 1024px resolution, Atlas-B achieves 91.04% accuracy, comparable to ConvNext-B (91.92%) while being 4.3x faster. Atlas is 2.95x faster and 7.38% better than FasterViT, 2.25x faster and 4.96% better than LongViT. In comparisons against MambaVision-S, we find Atlas-S achieves 5%, 16% and 32% higher accuracy at 1024px, 2048px and 4096px respectively, while obtaining similar runtimes. Code for reproducing our experiments and pretrained models is available at https://github.com/yalalab/atlas.