GUIDE: Guidance-based Incremental Learning with Diffusion Models
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910465078067200 |
|---|---|
| author | Cywiński, Bartosz Deja, Kamil Trzciński, Tomasz Twardowski, Bartłomiej Kuciński, Łukasz |
| author_facet | Cywiński, Bartosz Deja, Kamil Trzciński, Tomasz Twardowski, Bartłomiej Kuciński, Łukasz |
| contents | We introduce GUIDE, a novel continual learning approach that directs diffusion models to rehearse samples at risk of being forgotten. Existing generative strategies combat catastrophic forgetting by randomly sampling rehearsal examples from a generative model. Such an approach contradicts buffer-based approaches where sampling strategy plays an important role. We propose to bridge this gap by incorporating classifier guidance into the diffusion process to produce rehearsal examples specifically targeting information forgotten by a continuously trained model. This approach enables the generation of samples from preceding task distributions, which are more likely to be misclassified in the context of recently encountered classes. Our experimental results show that GUIDE significantly reduces catastrophic forgetting, outperforming conventional random sampling approaches and surpassing recent state-of-the-art methods in continual learning with generative replay. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2403_03938 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | GUIDE: Guidance-based Incremental Learning with Diffusion Models Cywiński, Bartosz Deja, Kamil Trzciński, Tomasz Twardowski, Bartłomiej Kuciński, Łukasz Machine Learning We introduce GUIDE, a novel continual learning approach that directs diffusion models to rehearse samples at risk of being forgotten. Existing generative strategies combat catastrophic forgetting by randomly sampling rehearsal examples from a generative model. Such an approach contradicts buffer-based approaches where sampling strategy plays an important role. We propose to bridge this gap by incorporating classifier guidance into the diffusion process to produce rehearsal examples specifically targeting information forgotten by a continuously trained model. This approach enables the generation of samples from preceding task distributions, which are more likely to be misclassified in the context of recently encountered classes. Our experimental results show that GUIDE significantly reduces catastrophic forgetting, outperforming conventional random sampling approaches and surpassing recent state-of-the-art methods in continual learning with generative replay. |
| title | GUIDE: Guidance-based Incremental Learning with Diffusion Models |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2403.03938 |