CULT: Continual Unsupervised Learning with Typicality-Based Environment Detection

Oliver Daniels-Koch
7/17/2022
cs.LG

No confident code implementation found for this paper.

Abstract

We introduce CULT (Continual Unsupervised Representation Learning with Typicality-Based Environment Detection), a new algorithm for continual unsupervised learning with variational auto-encoders. CULT uses a simple typicality metric in the latent space of a VAE to detect distributional shifts in the environment, which is used in conjunction with generative replay and an auxiliary environmental classifier to limit catastrophic forgetting in unsupervised representation learning. In our experiments, CULT significantly outperforms baseline continual unsupervised learning approaches. Code for this paper can be found here: https://github.com/oliveradk/cult

View on arXivView PDF

Code Implementations

No confident code match yet

We couldn't find an author-owned or strongly-evidenced community implementation for this paper. Any repos shown below are weak matches — verify before relying on them.

No code implementations found yet.

Know of an implementation? Let us know in the comments below!

Similar Papers

On the Effectiveness of LayerNorm Tuning for Continual Learning in Vision Transformers

83% match1 repo82023

Variational Self-Supervised Learning

82% match2025

Class Incremental Continual Learning with Self-Organizing Maps and Variational Autoencoders Using Synthetic Replay

82% match2025

FPPL: An Efficient and Non-IID Robust Federated Continual Learning Framework

82% match1 repo52024

Infinite dSprites for Disentangled Continual Learning: Separating Memory Edits from Generalization

82% match2 repos72023

Out-of-Distribution Detection in Continual Learning

82% match2026

Cite this paper

@article{danielskoch2022cult,
  title  = {CULT: Continual Unsupervised Learning with Typicality-Based Environment Detection},
  author = {Oliver Daniels-Koch},
  year   = {2022},
  eprint = {2207.08309},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2207.08309v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2207.08309.svg)](https://codeofpaper.com/paper/2207.08309)

Discussion