Differentiate-and-Inject: Enhancing VLAs via Functional Differentiation Induced by In-Parameter Structural Reasoning

TL;DR

This paper introduces iSTAR, a framework that embeds task-level semantic structures directly into vision-language-action models, improving their reasoning, generalization, and success rates in robotic manipulation tasks.

Contribution

iSTAR innovatively incorporates in-parameter structural reasoning to enhance VLA models, enabling better task decomposition without external planners or prompts.

Findings

Achieves higher success rates than baseline models.

Provides more reliable task decompositions.

Demonstrates effective generalization across diverse tasks.

Abstract

As robots are expected to perform increasingly diverse tasks, they must understand not only low-level actions but also the higher-level structure that determines how a task should unfold. Existing vision-language-action (VLA) models struggle with this form of task-level reasoning. They either depend on prompt-based in-context decomposition, which is unstable and sensitive to linguistic variations, or end-to-end long-horizon training, which requires large-scale demonstrations and entangles task-level reasoning with low-level control. We present in-parameter structured task reasoning (iSTAR), a framework for enhancing VLA models via functional differentiation induced by in-parameter structural reasoning. Instead of treating VLAs as monolithic policies, iSTAR embeds task-level semantic structure directly into model parameters, enabling differentiated task-level inference without external planners or handcrafted prompt inputs. This injected structure takes the form of implicit dynamic scene-graph knowledge that captures object relations, subtask semantics, and task-level dependencies in parameter space. Across diverse manipulation benchmarks, iSTAR achieves more reliable task decompositions and higher success rates than both in-context and end-to-end VLA baselines, demonstrating the effectiveness of parameter-space structural reasoning for functional differentiation and improved generalization across task variations.