Skip to main navigation
Skip to search
Skip to main content
ÉTS Research Discovery Portal Home
Help & FAQ
Link opens in a new tab
Français
English
Search content at ÉTS Research Discovery Portal
Home
Profiles
Research units
Research output
Datasets
Press/Media
A Universal Source-Free Class Unlearning Framework via Synthetic Embeddings
Zahra Dehghani
, Pablo Piantanida
,
Mohammadhadi Shateri
École de technologie supérieure
Systems Engineering Department
LIVIA - Imaging, Vision and Artificial Intelligence Laboratory
École de technologie supérieure
Université Paris-Saclay
Research output
:
Contribution to journal
›
Journal Article
›
peer-review
Overview
Fingerprint
Fingerprint
Dive into the research topics of 'A Universal Source-Free Class Unlearning Framework via Synthetic Embeddings'. Together they form a unique fingerprint.
Sort by
Weight
Alphabetically
Keyphrases
Source-free
100%
Unlearning
100%
Neural Network Classifier
20%
Generated Samples
20%
Logit
20%
Minimal Impact
10%
Computationally Expensive
10%
Source Data
10%
Benchmark Dataset
10%
Intermediate Space
10%
CIFAR-10
10%
Decision Boundary
10%
CIFAR-100
10%
ResNet50
10%
Effective Source
10%
ResNet18
10%
Backbone Architecture
10%
Ethical Requirements
10%
Swin-T
10%
Legal Requirements
10%
Parametric Form
10%
Novel Sources
10%
Embedded Distribution
10%
Computer Science
Residual Neural Network
100%
Experimental Result
50%
Effective Support
50%
Decision Boundary
50%
Vertebral Column
50%
Parametric Form
50%