Aghasanli, Agil and Angelov, Plamen (2026) Pull-to-Outlier & Contrastive Objective-level (POCO) Unlearning : A Framework for Sample and Objective Forgetting. Transactions on Machine Learning Research, 2026. ISSN 2835-8856
Full text not available from this repository.Abstract
Current Machine Unlearning (MU) methods require full retraining or extensive fine-tuning, lack formal removal criteria, and focus only on sample-level forgetting, limiting their practicality. We address these gaps with two lightweight, projection-only techniques operating above frozen feature extractors. Pull-to-Outlier Unlearning (POU) offers a transparent, unsupervised geometric removal method by displacing embeddings of unwanted samples or entire classes into synthetic outlier regions, while preserving downstream performance and distilling knowledge of the remaining data. To the best of our knowledge, Contrastive Objective-level Unlearning (COU) is the first method to remove learned objectives. It perturbs projection weights to eliminate a target objective’s influence. Then it realigns the original data manifold after objective perturbation. We validate POU on CIFAR10, CIFAR100, and Caltech-256 with ResNet-based backbones, showing efficient instance and class forgetting with minimal impact on retained accuracy. COU is tested on DINO and CLIP feature representations, providing evidence for objective-level removal in the studied settings while preserving non-target structure.