GLUE: Global-Local Unified Encoding for Imitation Learning via Key-Patch Tracking  

Anonymous ICRA Submission

Abstract

In recent years, visual representation learning has gained widespread attention in robotic imitation learning. However, in complex Out-of-Distribution(OOD) settings characterized by clutter and occlusion, the attention of global visual representations can be diluted or interfered, leading to degraded policy performance. The invariance of local representations for task-relevant objects offers a solution. By efficiently utilizing these local representations, training and testing data can be mapped to a more similar feature space, thereby mitigating the covariate shift problem. Accordingly, we propose GLUE, a global-local unified encoding framework for imitation learning based on key-patch tracking. GLUE selects and tracks key-patches as critical local representations by employing a text-guided mechanism. It features a novel fusion framework where global patch features query local patches to distill essential information, yielding fine-grained local features with low heterogeneity relative to the global context. This fused representation steers the robot’s visual attention toward task-relevant objects and preserves precise global context, which together align the training and testing distributions into a similar and task-informative feature space, ultimately enhancing the robustness of the imitation learning policy. Experiments demonstrate that GLUE achieves strong performance across diverse tasks in both simulation and realworld settings, outperforming the strongest baseline by 17.6% in simulation, 36.3% in real-world environments, and 58.3% on real-world generalization settings.


Evaluation Results in In-Domain Environments

Push Button(DP)

Push Button(ACT)

Push Button(GLUE-S)

Push Button(GLUE)

Stack Block(DP)

Stack Block(ACT)

Stack Block(GLUE-S)

Stack Block(GLUE)

Place Fruit(DP)

Place Fruit(ACT)

Place Fruit(GLUE-S)

Place Fruit(GLUE)

Fold Towel(DP)

Fold Towel(ACT)

Fold Towel(GLUE-S)

Fold Towel(GLUE)


Evaluation Results in Cluttered OOD Environments

Push Button(DP)

Push Button(ACT)

Push Button(GLUE-S)

Push Button(GLUE)

Stack Block(DP)

Stack Block(ACT)

Stack Block(GLUE-S)

Stack Block(GLUE)


Evaluation Results in Occluded OOD Environments

Push Button(DP)

Push Button(ACT)

Push Button(GLUE-S)

Push Button(GLUE)

Place Fruit(DP)

Place Fruit(ACT)

Place Fruit(GLUE-S)

Place Fruit(GLUE)


Evaluation Results in Illumination-Disturbed OOD Environments

Place Fruit(DP)

Place Fruit(ACT)

Place Fruit(GLUE-S)

Place Fruit(GLUE)

Fold Towel(DP)

Fold Towel(ACT)

Fold Towel(GLUE-S)

Fold Towel(GLUE)