Spatial Representation Distillation and Knowledge Routing for Vision-Language-Action Models
Overview
This paper introduces SPARK-VLA, a framework designed to enhance vision-language-action models (VLAs) with three-dimensional spatial understanding without relying on external geometric inputs. The method employs two main components: spatial representation distillation, which transfers 3D knowledge from a frozen expert to a LoRA-adapted visual branch while maintaining the original visual pathway, and a lightweight knowledge router that directs action-relevant visual tokens to the language model based on gradients from action loss. Evaluated on benchmarks including LIBERO, LIBERO-Plus, CALVIN, and real-world manipulation tasks, SPARK-VLA demonstrates strong performance using only a 0.5B parameter language backbone and approximately 53M trainable parameters. The code is publicly available.
Generated from attributed reports · Updated 12 hours ago
Event evidence and corrections
0 attributed source owners. Ownership does not establish independent confirmation. Quantities are reported separately and are never added together.
Artifact availability · code: available
“to 3D-aware VLA learning. Code is available at https://github.com/park-jungin/spark-vla.”
Exact source · revision 1Source owner not reported
Report timeline
Follow attributed reports and material updates.
- arXiv Robotics — research abstractsSignalSpatial Representation Distillation and Knowledge Routing for Vision-Language-Action Models
This paper introduces SPARK-VLA, a framework designed to enhance vision-language-action models (VLAs) with three-dimensional spatial understanding without relying on external geometric inputs. The method employs two main components: spatial representation distillation, which transfers 3D knowledge from a frozen expert to a LoRA-adapted visual branch while maintaining the original visual pathway, and a lightweight knowledge router that directs action-relevant visual tokens to the language model based on gradients from action loss. Evaluated on benchmarks including LIBERO, LIBERO-Plus, CALVIN, and real-world manipulation tasks, SPARK-VLA demonstrates strong performance using only a 0.5B parameter language backbone and approximately 53M trainable parameters. The code is publicly available.
Event coverage history
There is not enough continuous observation data to show a trend.
Timezone · UTC
Article dates follow your selected timezone. Briefing editions use Hong Kong time (UTC+8).