Unlearning via Sparse Representations

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Shah, Vedant, Träuble, Frederik, Malik, Ashish, Larochelle, Hugo, Mozer, Michael, Arora, Sanjeev, Bengio, Yoshua, Goyal, Anirudh
Format: Preprint
Veröffentlicht: 2023
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866913541154406400
author Shah, Vedant
Träuble, Frederik
Malik, Ashish
Larochelle, Hugo
Mozer, Michael
Arora, Sanjeev
Bengio, Yoshua
Goyal, Anirudh
author_facet Shah, Vedant
Träuble, Frederik
Malik, Ashish
Larochelle, Hugo
Mozer, Michael
Arora, Sanjeev
Bengio, Yoshua
Goyal, Anirudh
contents Machine \emph{unlearning}, which involves erasing knowledge about a \emph{forget set} from a trained model, can prove to be costly and infeasible by existing techniques. We propose a nearly compute-free zero-shot unlearning technique based on a discrete representational bottleneck. We show that the proposed technique efficiently unlearns the forget set and incurs negligible damage to the model's performance on the rest of the data set. We evaluate the proposed technique on the problem of \textit{class unlearning} using three datasets: CIFAR-10, CIFAR-100, and LACUNA-100. We compare the proposed technique to SCRUB, a state-of-the-art approach which uses knowledge distillation for unlearning. Across all three datasets, the proposed technique performs as well as, if not better than SCRUB while incurring almost no computational cost.
format Preprint
id arxiv_https___arxiv_org_abs_2311_15268
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Unlearning via Sparse Representations
Shah, Vedant
Träuble, Frederik
Malik, Ashish
Larochelle, Hugo
Mozer, Michael
Arora, Sanjeev
Bengio, Yoshua
Goyal, Anirudh
Machine Learning
Artificial Intelligence
Machine \emph{unlearning}, which involves erasing knowledge about a \emph{forget set} from a trained model, can prove to be costly and infeasible by existing techniques. We propose a nearly compute-free zero-shot unlearning technique based on a discrete representational bottleneck. We show that the proposed technique efficiently unlearns the forget set and incurs negligible damage to the model's performance on the rest of the data set. We evaluate the proposed technique on the problem of \textit{class unlearning} using three datasets: CIFAR-10, CIFAR-100, and LACUNA-100. We compare the proposed technique to SCRUB, a state-of-the-art approach which uses knowledge distillation for unlearning. Across all three datasets, the proposed technique performs as well as, if not better than SCRUB while incurring almost no computational cost.
title Unlearning via Sparse Representations
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2311.15268