GET-Zero: Graph Embodiment Transformer for Zero-shot Embodiment Generalization
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_ | 1866913495246700544 |
|---|---|
| author | Patel, Austin Song, Shuran |
| author_facet | Patel, Austin Song, Shuran |
| contents | This paper introduces GET-Zero, a model architecture and training procedure for learning an embodiment-aware control policy that can immediately adapt to new hardware changes without retraining. To do so, we present Graph Embodiment Transformer (GET), a transformer model that leverages the embodiment graph connectivity as a learned structural bias in the attention mechanism. We use behavior cloning to distill demonstration data from embodiment-specific expert policies into an embodiment-aware GET model that conditions on the hardware configuration of the robot to make control decisions. We conduct a case study on a dexterous in-hand object rotation task using different configurations of a four-fingered robot hand with joints removed and with link length extensions. Using the GET model along with a self-modeling loss enables GET-Zero to zero-shot generalize to unseen variation in graph structure and link length, yielding a 20% improvement over baseline methods. All code and qualitative video results are on https://get-zero-paper.github.io |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2407_15002 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | GET-Zero: Graph Embodiment Transformer for Zero-shot Embodiment Generalization Patel, Austin Song, Shuran Robotics I.2.9 This paper introduces GET-Zero, a model architecture and training procedure for learning an embodiment-aware control policy that can immediately adapt to new hardware changes without retraining. To do so, we present Graph Embodiment Transformer (GET), a transformer model that leverages the embodiment graph connectivity as a learned structural bias in the attention mechanism. We use behavior cloning to distill demonstration data from embodiment-specific expert policies into an embodiment-aware GET model that conditions on the hardware configuration of the robot to make control decisions. We conduct a case study on a dexterous in-hand object rotation task using different configurations of a four-fingered robot hand with joints removed and with link length extensions. Using the GET model along with a self-modeling loss enables GET-Zero to zero-shot generalize to unseen variation in graph structure and link length, yielding a 20% improvement over baseline methods. All code and qualitative video results are on https://get-zero-paper.github.io |
| title | GET-Zero: Graph Embodiment Transformer for Zero-shot Embodiment Generalization |
| topic | Robotics I.2.9 |
| url | https://arxiv.org/abs/2407.15002 |