BiSSL: Enhancing the Alignment Between Self-Supervised Pretraining and Downstream Fine-Tuning via Bilevel Optimization
Fuente:
arXiv
Guardado en:
| Autores principales: | , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866917261426556928 |
|---|---|
| author | Zakarias, Gustav Wagner Hansen, Lars Kai Tan, Zheng-Hua |
| author_facet | Zakarias, Gustav Wagner Hansen, Lars Kai Tan, Zheng-Hua |
| contents | Models initialized from self-supervised pretraining may suffer from poor alignment with downstream tasks, reducing the extent to which subsequent fine-tuning can adapt pretrained features toward downstream objectives. To mitigate this, we introduce BiSSL, a novel bilevel training framework that enhances the alignment of self-supervised pretrained models with downstream tasks prior to fine-tuning. BiSSL acts as an intermediate training stage conducted after conventional self-supervised pretraining and is tasked with solving a bilevel optimization problem that incorporates the pretext and downstream training objectives in its lower- and upper-level objectives, respectively. This approach explicitly models the interdependence between the pretraining and fine-tuning stages within the conventional self-supervised learning pipeline, facilitating enhanced information sharing between them that ultimately leads to a model initialization better aligned with the downstream task. We propose a general training algorithm for BiSSL that is compatible with a broad range of pretext and downstream tasks. Using SimCLR and Bootstrap Your Own Latent to pretrain ResNet-50 backbones on the ImageNet dataset, we demonstrate that our proposed framework significantly improves accuracy on the vast majority of 12 downstream image classification datasets, as well as on object detection. Exploratory analyses alongside investigative experiments further provide compelling evidence that BiSSL enhances downstream alignment. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_02387 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | BiSSL: Enhancing the Alignment Between Self-Supervised Pretraining and Downstream Fine-Tuning via Bilevel Optimization Zakarias, Gustav Wagner Hansen, Lars Kai Tan, Zheng-Hua Machine Learning Artificial Intelligence Models initialized from self-supervised pretraining may suffer from poor alignment with downstream tasks, reducing the extent to which subsequent fine-tuning can adapt pretrained features toward downstream objectives. To mitigate this, we introduce BiSSL, a novel bilevel training framework that enhances the alignment of self-supervised pretrained models with downstream tasks prior to fine-tuning. BiSSL acts as an intermediate training stage conducted after conventional self-supervised pretraining and is tasked with solving a bilevel optimization problem that incorporates the pretext and downstream training objectives in its lower- and upper-level objectives, respectively. This approach explicitly models the interdependence between the pretraining and fine-tuning stages within the conventional self-supervised learning pipeline, facilitating enhanced information sharing between them that ultimately leads to a model initialization better aligned with the downstream task. We propose a general training algorithm for BiSSL that is compatible with a broad range of pretext and downstream tasks. Using SimCLR and Bootstrap Your Own Latent to pretrain ResNet-50 backbones on the ImageNet dataset, we demonstrate that our proposed framework significantly improves accuracy on the vast majority of 12 downstream image classification datasets, as well as on object detection. Exploratory analyses alongside investigative experiments further provide compelling evidence that BiSSL enhances downstream alignment. |
| title | BiSSL: Enhancing the Alignment Between Self-Supervised Pretraining and Downstream Fine-Tuning via Bilevel Optimization |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2410.02387 |