Search papers, labs, and topics across Lattice.
This paper investigates class unlearning in a source-free context, focusing on the recoverability of forget classes through a classifier-head update without requiring additional data. By establishing a theoretical alignment condition, the authors introduce the Source-Free Relearning Audit (SFRA), which generates candidate embeddings to relabel probes as forget classes while maintaining accuracy on retain classes. Experiments demonstrate that several unlearning methods can achieve significant source-free recoverability, with some methods outperforming traditional retraining benchmarks on datasets like CIFAR-10 and TinyImageNet.
Forget classes can be effectively recovered in a source-free setting, with some methods even surpassing traditional retraining performance.
Class unlearning aims to remove a model's ability to recognize designated forget classes while preserving performance on retain classes. However, low forget accuracy after unlearning does not necessarily mean the class structure has been erased. Approximate unlearning methods can alter classifier decision boundaries while leaving recoverable structure in the representation. Prior work has shown that forget classes can be recovered, but existing approaches require real forget or retain samples, auxiliary data, or reference checkpoints. We study class relearning in a strictly source-free setting, asking whether a forget class can be recovered through a classifier-head update using only the unlearned model. Our approach rests on a theoretical analysis establishing a sufficient alignment condition under which a single gradient step on a synthetic probe set increases the expected logit margin of the forget class. Building on this, we propose a white-box Source-Free Relearning Audit (SFRA), which generates candidate embeddings in representation space and uses model-guided confidence filtering to construct high-confidence retain probes and low-confidence boundary-adjacent probes that are relabelled as the forget class. Gaussian sampling and Softmax confidence are used by default, while ablations with alternative proposal distributions and uncertainty criteria show that recoverability is not specific to these choices. To quantify recoverability, we introduce the Relearning Score (RS), which jointly measures forget-class recovery and retain-accuracy preservation, and report class-matched $\Delta$RS relative to a retrained reference. Experiments on CIFAR-10, CIFAR-100, and TinyImageNet with ResNet-18, ViT-B/16, and Swin-T show that several unlearning methods exhibit substantial source-free recoverability, and that for a subset of methods this recoverability exceeds the matched retrained reference.
