Learning to Move Before Learning to Do: Task-Agnostic pretraining for VLAs
Task-Agnostic Pretraining framework trains robotic models using self-supervised inverse dynamics on unlabeled data followed by lightweight language grounding, achieving superior pe…
Hugging Face · Daily Papers
·Junhao Shi, Siyin Wang
·
·▲ 4 upvotes
Este artigo está em destaque na seleção diária de papers do Hugging Face, curada pela comunidade de pesquisa em IA.
Autores: Junhao Shi, Siyin Wang, Xiaopeng Yu, Li Ji, Jingjing Gong, Xipeng Qiu
- 4 upvotes da comunidade
- Temas: Vision-Language-Action models, expert demonstrations, physical competence, semantic alignment, self-supervised Inverse Dynamics, task-agnostic pretraining
Resumo
Resumo original (em inglês), extraído do paper:
Task-Agnostic Pretraining framework trains robotic models using self-supervised inverse dynamics on unlabeled data followed by lightweight language grounding, achieving superior performance with minimal expert demonstrations.Onde ler
// relacionados
Leia também
Blog
Orientação de Segurança Neuro-Simbólica para Modelos de Visão-Linguagem-Ação via Correspondência de Fluxo Restrita
Blog
PairCoder++: Pair Programming as a Universal Paradigm for Verified Code-Driven Multimodal and Structured-Artifact Generation
Blog
Guided Action Flow: Q-Guided Inference for Flow-Matching Vision-Language-Action Policies
Blog