ActPrune: Action-Grounded Visual Token Pruning for Efficient Vision-Language-Action Models
cs.RO
Submitted: 2026-11-06
Updated: 2026-11-11
Project page: http://mmai.ewha.ac.kr/actprune
Terminology
Sources
- RoboMamba: Efficient Vision-Language-Action Model for Robotic Reasoning and Manipulation
- TinyVLA: Towards Fast, Data-Efficient Vision-Language-Action Models for Robotic Manipulation
- Quantization-Aware Imitation-Learning for Resource-Efficient Robotic Control
- Bridging the Semantic-Action Gap in Visual Token Pruning for Efficient VLA Inference
- FAST: Efficient Action Tokenization for Vision-Language-Action Models
- Llama 2: Open Foundation and Fine-Tuned Chat Models
- Gemma 2: Improving Open Language Models at a Practical Size
- DINOv2: Learning Robust Visual Features without Supervision
Related papers
- Rethinking Actions: Continuity-Aware Priors for World-Model Planning
- MOSAIC: Modular Obstacle and Spatial-Aware Framework for Manipulation In Cluttered Scenes
- SAGAS: Semantic-Aware Graph-Assisted Stitching for Offline Temporal Logic Planning
- DexSeed: Expanding Human Hand-Object Interactions into Robot Demonstrations for Dexterous Policy Learning
- LaST0.5: A Generalist World Action Model with Interaction-Centric Latent Reasoning
- CIALLO: Learning Shared Intent Representations for Cross-Embodiment Manipulation