# Generative Myopia: Why Diffusion Models Fail at Structure

arXiv: https://arxiv.org/abs/2511.18593 · HTML: https://mlsanity.com/paper/2511.18593/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Diffusion foundations and sampling, one of 14 directions in Diffusion models.
- Builds on 7 of its 11 references that are in the map.
- First posted on arXiv 23 Nov 2025.

## Placed in

- Diffusion models / [Diffusion foundations and sampling](https://mlsanity.com/diffusion/diffusion-foundations-and-sampling/)

## Builds on (most cited first, 7 of 7)

- [Denoising Diffusion Probabilistic Models](https://mlsanity.com/paper/2006.11239/) (2020)
- [ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness](https://mlsanity.com/paper/1811.12231/) (2018)
- [On the Spectral Bias of Neural Networks](https://mlsanity.com/paper/1806.08734/) (2018)
- [DIFUSCO: Graph-based Diffusion Solvers for Combinatorial Optimization](https://mlsanity.com/paper/2302.08224/) (2023)
- [Fast Graph Generation via Spectral Diffusion](https://mlsanity.com/paper/2211.08892/) (2022)
- [Do Graph Diffusion Models Accurately Capture and Generate Substructure Distributions?](https://mlsanity.com/paper/2502.02488/) (2025)
- [Spectral Neural Graph Sparsification](https://mlsanity.com/paper/2510.27474/) (2025)

## Abstract (arXiv)

Graph Diffusion Models (GDMs) optimize for statistical likelihood, implicitly acting as frequency filters that favor abundant substructures over spectrally critical ones. We term this phenomenon Generative Myopia. In combinatorial tasks like graph sparsification, this leads to the catastrophic removal of ``rare bridges,'' edges that are structurally mandatory (Reff ≈ 1) but statistically scarce. We prove theoretically and empirically that this failure is driven by Gradient Starvation: the optimization landscape itself suppresses rare structural signals, rendering them unlearnable regardless of model capacity. To resolve this, we introduce Spectrally-Weighted Diffusion, which re-aligns the variational objective using Effective Resistance. We demonstrate that spectral priors can be amortized into the training phase with zero inference overhead. Our method eliminates myopia, matching the performance of an optimal Spectral Oracle and achieving 100% connectivity on adversarial benchmarks where standard diffusion fails completely (0%).