Model Adaptation for Time Constrained Embodied Control
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_ | 1866929387983601664 |
|---|---|
| author | Song, Jaehyun Yoo, Minjong Woo, Honguk |
| author_facet | Song, Jaehyun Yoo, Minjong Woo, Honguk |
| contents | When adopting a deep learning model for embodied agents, it is required that the model structure be optimized for specific tasks and operational conditions. Such optimization can be static such as model compression or dynamic such as adaptive inference. Yet, these techniques have not been fully investigated for embodied control systems subject to time constraints, which necessitate sequential decision-making for multiple tasks, each with distinct inference latency limitations. In this paper, we present MoDeC, a time constraint-aware embodied control framework using the modular model adaptation. We formulate model adaptation to varying operational conditions on resource and time restrictions as dynamic routing on a modular network, incorporating these conditions as part of multi-task objectives. Our evaluation across several vision-based embodied environments demonstrates the robustness of MoDeC, showing that it outperforms other model adaptation methods in both performance and adherence to time constraints in robotic manipulation and autonomous driving applications |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_11128 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Model Adaptation for Time Constrained Embodied Control Song, Jaehyun Yoo, Minjong Woo, Honguk Machine Learning Robotics When adopting a deep learning model for embodied agents, it is required that the model structure be optimized for specific tasks and operational conditions. Such optimization can be static such as model compression or dynamic such as adaptive inference. Yet, these techniques have not been fully investigated for embodied control systems subject to time constraints, which necessitate sequential decision-making for multiple tasks, each with distinct inference latency limitations. In this paper, we present MoDeC, a time constraint-aware embodied control framework using the modular model adaptation. We formulate model adaptation to varying operational conditions on resource and time restrictions as dynamic routing on a modular network, incorporating these conditions as part of multi-task objectives. Our evaluation across several vision-based embodied environments demonstrates the robustness of MoDeC, showing that it outperforms other model adaptation methods in both performance and adherence to time constraints in robotic manipulation and autonomous driving applications |
| title | Model Adaptation for Time Constrained Embodied Control |
| topic | Machine Learning Robotics |
| url | https://arxiv.org/abs/2406.11128 |