SelfReplay: Adapting Self-Supervised Sensory Models via Adaptive Meta-Task Replay
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_ | 1866916657454120960 |
|---|---|
| author | Yoon, Hyungjun Kwak, Jaehyun Tolera, Biniyam Aschalew Dai, Gaole Li, Mo Gong, Taesik Lee, Kimin Lee, Sung-Ju |
| author_facet | Yoon, Hyungjun Kwak, Jaehyun Tolera, Biniyam Aschalew Dai, Gaole Li, Mo Gong, Taesik Lee, Kimin Lee, Sung-Ju |
| contents | Self-supervised learning has emerged as a method for utilizing massive unlabeled data for pre-training models, providing an effective feature extractor for various mobile sensing applications. However, when deployed to end-users, these models encounter significant domain shifts attributed to user diversity. We investigate the performance degradation that occurs when self-supervised models are fine-tuned in heterogeneous domains. To address the issue, we propose SelfReplay, a few-shot domain adaptation framework for personalizing self-supervised models. SelfReplay proposes self-supervised meta-learning for initial model pre-training, followed by a user-side model adaptation by replaying the self-supervision with user-specific data. This allows models to adjust their pre-trained representations to the user with only a few samples. Evaluation with four benchmarks demonstrates that SelfReplay outperforms existing baselines by an average F1-score of 8.8%p. Our on-device computational overhead analysis on a commodity off-the-shelf (COTS) smartphone shows that SelfReplay completes adaptation within an unobtrusive latency (in three minutes) with only a 9.54% memory consumption, demonstrating the computational efficiency of the proposed method. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2404_15305 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | SelfReplay: Adapting Self-Supervised Sensory Models via Adaptive Meta-Task Replay Yoon, Hyungjun Kwak, Jaehyun Tolera, Biniyam Aschalew Dai, Gaole Li, Mo Gong, Taesik Lee, Kimin Lee, Sung-Ju Signal Processing Machine Learning Self-supervised learning has emerged as a method for utilizing massive unlabeled data for pre-training models, providing an effective feature extractor for various mobile sensing applications. However, when deployed to end-users, these models encounter significant domain shifts attributed to user diversity. We investigate the performance degradation that occurs when self-supervised models are fine-tuned in heterogeneous domains. To address the issue, we propose SelfReplay, a few-shot domain adaptation framework for personalizing self-supervised models. SelfReplay proposes self-supervised meta-learning for initial model pre-training, followed by a user-side model adaptation by replaying the self-supervision with user-specific data. This allows models to adjust their pre-trained representations to the user with only a few samples. Evaluation with four benchmarks demonstrates that SelfReplay outperforms existing baselines by an average F1-score of 8.8%p. Our on-device computational overhead analysis on a commodity off-the-shelf (COTS) smartphone shows that SelfReplay completes adaptation within an unobtrusive latency (in three minutes) with only a 9.54% memory consumption, demonstrating the computational efficiency of the proposed method. |
| title | SelfReplay: Adapting Self-Supervised Sensory Models via Adaptive Meta-Task Replay |
| topic | Signal Processing Machine Learning |
| url | https://arxiv.org/abs/2404.15305 |