Deep Classifier Mimicry without Data Access

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Braun, Steven, Mundt, Martin, Kersting, Kristian
Format: Preprint
Veröffentlicht: 2023
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866914770733498368
author Braun, Steven
Mundt, Martin
Kersting, Kristian
author_facet Braun, Steven
Mundt, Martin
Kersting, Kristian
contents Access to pre-trained models has recently emerged as a standard across numerous machine learning domains. Unfortunately, access to the original data the models were trained on may not equally be granted. This makes it tremendously challenging to fine-tune, compress models, adapt continually, or to do any other type of data-driven update. We posit that original data access may however not be required. Specifically, we propose Contrastive Abductive Knowledge Extraction (CAKE), a model-agnostic knowledge distillation procedure that mimics deep classifiers without access to the original data. To this end, CAKE generates pairs of noisy synthetic samples and diffuses them contrastively toward a model's decision boundary. We empirically corroborate CAKE's effectiveness using several benchmark datasets and various architectural choices, paving the way for broad application.
format Preprint
id arxiv_https___arxiv_org_abs_2306_02090
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Deep Classifier Mimicry without Data Access
Braun, Steven
Mundt, Martin
Kersting, Kristian
Machine Learning
Artificial Intelligence
Access to pre-trained models has recently emerged as a standard across numerous machine learning domains. Unfortunately, access to the original data the models were trained on may not equally be granted. This makes it tremendously challenging to fine-tune, compress models, adapt continually, or to do any other type of data-driven update. We posit that original data access may however not be required. Specifically, we propose Contrastive Abductive Knowledge Extraction (CAKE), a model-agnostic knowledge distillation procedure that mimics deep classifiers without access to the original data. To this end, CAKE generates pairs of noisy synthetic samples and diffuses them contrastively toward a model's decision boundary. We empirically corroborate CAKE's effectiveness using several benchmark datasets and various architectural choices, paving the way for broad application.
title Deep Classifier Mimicry without Data Access
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2306.02090