Search papers, labs, and topics across Lattice.
VLA-Pro is introduced, a plug-and-play framework that improves cross-task generalization in Vision-Language-Action (VLA) models by storing and transferring task-relevant procedural memories. During training, task-specific LoRA adapters are stored as parameterized procedural memories, which are then retrieved and fused at inference time based on the current multi-modal context to generate actions. Experiments across RoboTwin, RLBench, and real-world manipulation tasks demonstrate that VLA-Pro significantly enhances cross-task generalization, achieving up to a 207% relative improvement in simulation and boosting real-world success rates from 5.8% to 65.0%.
Forget fine-tuning: VLA-Pro dynamically fuses task-specific LoRA adapters retrieved from memory to achieve state-of-the-art cross-task generalization in robotic manipulation.
Vision-Language-Action~(VLA) models have shown strong potential for general-purpose robotic manipulation, yet they still struggle to generalize to unseen tasks that necessitate transferring relevant experience across objects, scenes, and action patterns. This paper proposes VLA-Pro, a plug-and-play framework designed to enhance cross-task generalization by storing task-relevant procedural memories at training time and transferring these memories during inference. Specifically, VLA-Pro stores task-specific LoRA adapters as parameterized procedural memories during training. At inference time, VLA-Pro retrieves relevant procedural memories based on the current multi-modal context and dynamically fuses these memories for generating the current action chunk. Experiments on RoboTwin, RLBench, and real-world manipulation tasks show that VLA-Pro consistently improves cross-task generalization across multiple backbones, achieving up to a 207% relative improvement in simulation and increasing real-world success rate from 5.8% to 65.0%. These results suggest that procedural memory retrieval and adaptation provide an effective mechanism for transferring manipulation experience to novel tasks while preserving modularity and execution stability.
