SLAM-AGS: Slide-Label Aware Multi-Task Pretraining Using Adaptive Gradient Surgery in Computational Cytology

TL;DR

SLAM-AGS introduces a multitask pretraining framework with adaptive gradient surgery for computational cytology, improving downstream performance especially at low witness rates by addressing unreliable labels and conflicting gradients.

Contribution

The paper presents SLAM-AGS, a novel pretraining method that combines weakly supervised and self-supervised objectives with adaptive gradient surgery for stability.

Findings

Enhanced bag-level F1-Score at low witness rates

Improved positive cell retrieval performance

Stable pretraining with conflicting gradient management

Abstract

Computational cytology faces two major challenges: i) instance-level labels are unreliable and prohibitively costly to obtain, ii) witness rates are extremely low. We propose SLAM-AGS, a Slide-Label-Aware Multitask pretraining framework that jointly optimizes (i) a weakly supervised similarity objective on slide-negative patches and (ii) a self-supervised contrastive objective on slide-positive patches, yielding stronger performance on downstream tasks. To stabilize learning, we apply Adaptive Gradient Surgery to tackle conflicting task gradients and prevent model collapse. We integrate the pretrained encoder into an attention-based Multiple Instance Learning aggregator for bag-level prediction and attention-guided retrieval of the most abnormal instances in a bag. On a publicly available bone-marrow cytology dataset, with simulated witness rates from 10% down to 0.5%, SLAM-AGS improves bag-level F1-Score and Top 400 positive cell retrieval over other pretraining methods, with the largest gains at low witness rates, showing that resolving gradient interference enables stable pretraining and better performance on downstream tasks. To facilitate reproducibility, we share our complete implementation and evaluation framework as open source: https://github.com/Ace95/SLAM-AGS.