Deep Unlearning via Randomized Conditionally Independent Hessians
This work addresses the need for efficient unlearning in deep learning models due to legal, privacy, or data corruption issues, offering a scalable solution for domains like face recognition and person re-identification.
The paper tackles the problem of machine unlearning for deep models by introducing L-CODEC, a method that identifies a subset of model parameters with high semantic overlap on a per-sample basis to avoid inverting large Hessian matrices, enabling approximate unlearning in previously infeasible settings like vision and NLP models.
Recent legislation has led to interest in machine unlearning, i.e., removing specific training samples from a predictive model as if they never existed in the training dataset. Unlearning may also be required due to corrupted/adversarial data or simply a user's updated privacy requirement. For models which require no training (k-NN), simply deleting the closest original sample can be effective. But this idea is inapplicable to models which learn richer representations. Recent ideas leveraging optimization-based updates scale poorly with the model dimension d, due to inverting the Hessian of the loss function. We use a variant of a new conditional independence coefficient, L-CODEC, to identify a subset of the model parameters with the most semantic overlap on an individual sample level. Our approach completely avoids the need to invert a (possibly) huge matrix. By utilizing a Markov blanket selection, we premise that L-CODEC is also suitable for deep unlearning, as well as other applications in vision. Compared to alternatives, L-CODEC makes approximate unlearning possible in settings that would otherwise be infeasible, including vision models used for face recognition, person re-identification and NLP models that may require unlearning samples identified for exclusion. Code can be found at https://github.com/vsingh-group/LCODEC-deep-unlearning/