# Conditioning noise is a free regularizer for LoRA fine-tuning: no pathology encoder required for diffusion-based artifact detection in histopathology

arXiv: https://arxiv.org/abs/2609.16032 · HTML: https://mlsanity.com/paper/2609.16032/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 3 of its 6 references that are in the map.
- First posted on arXiv 11 Sep 2026.

## Builds on (most cited first, 3 of 3)

- [LoRACLR: Contrastive Adaptation for Customization of Diffusion Models](https://mlsanity.com/paper/2412.09622/) (2024)
- [Correcting Deviations from Normality: A Reformulated Diffusion Model for Multi-Class Unsupervised Anomaly Detection](https://mlsanity.com/paper/2503.19357/) (2025)
- [Diffuse to Detect: Generative Diffusion Models for Unsupervised IC Anomaly Detection](https://mlsanity.com/paper/2605.26468/) (2026)

## Built on by (most cited first, 1 of 1)

- [The Effect of Tissue Detection on False Positives of Diffusion-Based Artifact Detection in Histopathology](https://mlsanity.com/paper/2609.40083/) (2026)

## Abstract (arXiv)

Diffusion-based artifact detectors score whole-slide image patches by reconstruction error under a model fine-tuned on clean tissue. We show that conditioning this fine-tuning on random Gaussian embeddings -- resampled at every step from approx. 200 KB of precomputed embedding statistics, with no encoder, no cache, and no change to inference -- consistently widens the clean/artifact separation. A four-step ablation chain shows the benefit requires neither content (shuffled real embeddings), provenance (synthetic Gaussians), a tuned intensity (flat across an 8x variance range), nor per-patch identity (fresh per-step noise); a LoRA-dropout control shows the conditioning pathway specifically, not generic weight perturbation, carries the effect. Patch-level gains of +0.25-0.48 Cohen's d replicate across nine trainings; honest leave-one-slide-out evaluation clears a pre-registered bar in 2/2 seeds; and two pre-registered external endpoints on a 281-case set confirm pooled Delta F1 = +0.0073 (95% CI) and +0.0129 (97.5% CI, two-look corrected). We release the full evaluation protocol, including measured seed noise and selection-optimism pricing.