# Indirect Attention: Turning Context Misalignment into a Feature

arXiv: https://arxiv.org/abs/2509.26015 · HTML: https://mlsanity.com/paper/2509.26015/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 31 of its 39 references that are in the map.
- First posted on arXiv 30 Sep 2025.

## Builds on (most cited first, 30 of 31)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://mlsanity.com/paper/1910.10683/) (2019)
- [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://mlsanity.com/paper/2103.14030/) (2021)
- [Intriguing properties of neural networks](https://mlsanity.com/paper/1312.6199/) (2013)
- [Flamingo: a Visual Language Model for Few-Shot Learning](https://mlsanity.com/paper/2204.14198/) (2022)
- [Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification](https://mlsanity.com/paper/1502.01852/) (2015)
- [Perceiver: General Perception with Iterative Attention](https://mlsanity.com/paper/2103.03206/) (2021)
- [Train Short, Test Long: Attention with Linear Biases Enables Input Length Extrapolation](https://mlsanity.com/paper/2108.12409/) (2021)
- [A Generalist Agent](https://mlsanity.com/paper/2205.06175/) (2022)
- [Fully-Convolutional Siamese Networks for Object Tracking](https://mlsanity.com/paper/1606.09549/) (2016)
- [Attention is Not All You Need: Pure Attention Loses Rank Doubly Exponentially with Depth](https://mlsanity.com/paper/2103.03404/) (2021)
- [Multimodal Foundation Models: From Specialists to General-Purpose Assistants](https://mlsanity.com/paper/2309.10020/) (2023)
- [Rethinking and Improving Relative Position Encoding for Vision Transformer](https://mlsanity.com/paper/2107.14222/) (2021)
- [Implicit regularization for deep neural networks driven by an Ornstein-Uhlenbeck like process](https://mlsanity.com/paper/1904.09080/) (2019)
- [Label Noise SGD Provably Prefers Flat Global Minimizers](https://mlsanity.com/paper/2106.06530/) (2021)
- [Shape Matters: Understanding the Implicit Bias of the Noise Covariance](https://mlsanity.com/paper/2006.08680/) (2020)
- [KERPLE: Kernelized Relative Positional Embedding for Length Extrapolation](https://mlsanity.com/paper/2205.09921/) (2022)
- [One-Shot Object Detection with Co-Attention and Co-Excitation](https://mlsanity.com/paper/1911.12529/) (2019)
- [Meta-DETR: Image-Level Few-Shot Detection with Inter-Class Correlation Exploitation](https://mlsanity.com/paper/2208.00219/) (2022)
- [One-Shot Instance Segmentation](https://mlsanity.com/paper/1811.11507/) (2018)
- [Improved Text Classification via Contrastive Adversarial Training](https://mlsanity.com/paper/2107.10137/) (2021)
- [FS-DETR: Few-Shot DEtection TRansformer with prompting and without re-training](https://mlsanity.com/paper/2210.04845/) (2022)
- [Alignment Attention by Matching Key and Query Distributions](https://mlsanity.com/paper/2110.12567/) (2021)
- [Semantic-aligned Fusion Transformer for One-shot Object Detection](https://mlsanity.com/paper/2203.09093/) (2022)
- [Benign Overfitting in Single-Head Attention](https://mlsanity.com/paper/2410.07746/) (2024)
- [Making Attention Mechanisms More Robust and Interpretable with Virtual Adversarial Training](https://mlsanity.com/paper/2104.08763/) (2021)
- [Multimodal Transformer Using Cross-Channel attention for Object Detection in Remote Sensing Images](https://mlsanity.com/paper/2310.13876/) (2023)

## Abstract (arXiv)

The attention mechanism has become a cornerstone of modern deep learning architectures, where keys and values are typically derived from the same underlying sequence or representation. This work explores a less conventional scenario, when keys and values originate from different sequences or modalities. Specifically, we first analyze the attention mechanism's behavior under noisy value features, establishing a critical noise threshold beyond which signal degradation becomes significant. Furthermore, we model context (key, value) misalignment as an effective form of structured noise within the value features, demonstrating that the noise induced by such misalignment can substantially exceed this critical threshold, thereby compromising standard attention's efficacy. Motivated by this, we introduce Indirect Attention, a modified attention mechanism that infers relevance indirectly in scenarios with misaligned context. We evaluate the performance of Indirect Attention across a range of synthetic tasks and real world applications, showcasing its superior ability to handle misalignment.