Track synthesis, materials and catalysis as it is published.
Search papers, labs, and topics across Lattice.
This paper introduces MorphAct, a contextual parameterization framework for robot policies that dynamically adjusts parameters based on interaction contexts to overcome the limitations of fixed shared parameters. By employing coupled generators that produce low-rank updates for vision-language and action modules, conditioned on computational dependencies, MorphAct enables more effective adaptation to diverse environments. Experimental results demonstrate significant performance improvements on benchmark datasets and real-world evaluations, highlighting the efficacy of context-dependent parameterization.
Robot policies can achieve up to 16% better performance by dynamically morphing their parameters with context, rather than relying on fixed, shared weights.
Progress toward general intelligence in the physical world requires scaling robot learning across diverse environments and interactions. However, existing robot policies use shared parameters that remain fixed across interaction contexts. When different operations favor competing parameter adjustments, fitting their mixture can leave individual operations insufficiently adapted. Our atomic-task diagnostics reveal competing update directions in both vision-language and action parameters, with source-improving updates increasing loss on other tasks. This motivates MorphAct, a contextual parameterization framework that dynamically adjusts robot policy parameters as interaction contexts change. Two coupled generators produce context-dependent low-rank updates for the vision-language and action modules. Their coupling follows the policy's computational dependency: vision-language updates shape the representations that condition action-side parameter generation. Both generators are trained jointly through the native action objective, without operation labels or expert-adapter targets. MorphAct improves performance across three VLA backbones on LIBERO and Meta-World, with gains of up to 16.30%. In real-world evaluation, it achieves gains of up to 14% over $\pi_{0.5}$. Ablations support the benefits of contextual parameter generation and representation coupling.
NUS