Year: 2022
Type: conference-paper
Abstract: 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 reidentification and NLP models that may require unlearning samples identified for exclusion. Code is available at https://github.com/vsingh-group/LCODEC-deep-unlearning
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
Awards: FMitF: Track I: Formal Methods for Explainable Machine Learning, Maximizing Power in AD Clinical Trials via Multimodal Machine Learning, Algebraic Formulations for Characterizing Structural Brain Connectivity Changes and Pathology Transmission Networks in Preclinical Alzheimer's Disease, Preclinical AD Consortium, CRCNS: Pooling Multisite data and Staistical Tests for analyzing biomarkers of Alzheimer's disease
OpenAlex ID: W4312595359
