This project investigates the idea of machine unlearning. It begins with examples on classification models and moves on to generative models like 🖼️ picture generators or Large Language Models (LLMs). Motivated by the movie Yesterday where the lead character wakes up in a Beatles-less world, this research explores the effects and techniques of successfully 'erasing' learned data from artificial intelligence models. The goal of the project is to create methods for removing particular data from trained models in a targeted manner without compromising overall performance. This is especially important in situations where maintaining data privacy and adhering to laws like the GDPR are crucial.
-
Notifications
You must be signed in to change notification settings - Fork 0
Exploring machine unlearning: from classification models to generative models like stable diffusion or LLMs. Inspired by 'Yesterday', it delves into removing learned data from AI models, aiming to preserve performance amid privacy and GDPR compliance challenges.
License
mich1803/Yesterday-Machine-Unlearning
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Folders and files
Name | Name | Last commit message | Last commit date | |
---|---|---|---|---|
Repository files navigation
About
Exploring machine unlearning: from classification models to generative models like stable diffusion or LLMs. Inspired by 'Yesterday', it delves into removing learned data from AI models, aiming to preserve performance amid privacy and GDPR compliance challenges.
Topics
Resources
License
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published