Enhancing Accuracy and Parameter-Efficiency of Neural Representations for Network Parameterization
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2024
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866914852542349312 |
|---|---|
| author | Choi, Hongjun Thiagarajan, Jayaraman J. Glatt, Ruben Liu, Shusen |
| author_facet | Choi, Hongjun Thiagarajan, Jayaraman J. Glatt, Ruben Liu, Shusen |
| contents | In this work, we investigate the fundamental trade-off regarding accuracy and parameter efficiency in the parameterization of neural network weights using predictor networks. We present a surprising finding that, when recovering the original model accuracy is the sole objective, it can be achieved effectively through the weight reconstruction objective alone. Additionally, we explore the underlying factors for improving weight reconstruction under parameter-efficiency constraints, and propose a novel training scheme that decouples the reconstruction objective from auxiliary objectives such as knowledge distillation that leads to significant improvements compared to state-of-the-art approaches. Finally, these results pave way for more practical scenarios, where one needs to achieve improvements on both model accuracy and predictor network parameter-efficiency simultaneously. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2407_00356 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Enhancing Accuracy and Parameter-Efficiency of Neural Representations for Network Parameterization Choi, Hongjun Thiagarajan, Jayaraman J. Glatt, Ruben Liu, Shusen Machine Learning Computer Vision and Pattern Recognition In this work, we investigate the fundamental trade-off regarding accuracy and parameter efficiency in the parameterization of neural network weights using predictor networks. We present a surprising finding that, when recovering the original model accuracy is the sole objective, it can be achieved effectively through the weight reconstruction objective alone. Additionally, we explore the underlying factors for improving weight reconstruction under parameter-efficiency constraints, and propose a novel training scheme that decouples the reconstruction objective from auxiliary objectives such as knowledge distillation that leads to significant improvements compared to state-of-the-art approaches. Finally, these results pave way for more practical scenarios, where one needs to achieve improvements on both model accuracy and predictor network parameter-efficiency simultaneously. |
| title | Enhancing Accuracy and Parameter-Efficiency of Neural Representations for Network Parameterization |
| topic | Machine Learning Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2407.00356 |