Attentive Feature Aggregation or: How Policies Learn to Stop Worrying about Robustness and Attend to Task-Relevant Visual Cues

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Tsagkas, Nikolaos, Sochopoulos, Andreas, Danier, Duolikun, Vijayakumar, Sethu, Kouris, Alexandros, Mac Aodha, Oisin, Lu, Chris Xiaoxuan
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866917299497205760
author Tsagkas, Nikolaos
Sochopoulos, Andreas
Danier, Duolikun
Vijayakumar, Sethu
Kouris, Alexandros
Mac Aodha, Oisin
Lu, Chris Xiaoxuan
author_facet Tsagkas, Nikolaos
Sochopoulos, Andreas
Danier, Duolikun
Vijayakumar, Sethu
Kouris, Alexandros
Mac Aodha, Oisin
Lu, Chris Xiaoxuan
contents The adoption of pre-trained visual representations (PVRs), leveraging features from large-scale vision models, has become a popular paradigm for training visuomotor policies. However, these powerful representations can encode a broad range of task-irrelevant scene information, making the resulting trained policies vulnerable to out-of-domain visual changes and distractors. In this work we address visuomotor policy feature pooling as a solution to the observed lack of robustness in perturbed scenes. We achieve this via Attentive Feature Aggregation (AFA), a lightweight, trainable pooling mechanism that learns to naturally attend to task-relevant visual cues, ignoring even semantically rich scene distractors. Through extensive experiments in both simulation and the real world, we demonstrate that policies trained with AFA significantly outperform standard pooling approaches in the presence of visual perturbations, without requiring expensive dataset augmentation or fine-tuning of the PVR. Our findings show that ignoring extraneous visual information is a crucial step towards deploying robust and generalisable visuomotor policies. Project Page: tsagkas.github.io/afa
format Preprint
id arxiv_https___arxiv_org_abs_2511_10762
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Attentive Feature Aggregation or: How Policies Learn to Stop Worrying about Robustness and Attend to Task-Relevant Visual Cues
Tsagkas, Nikolaos
Sochopoulos, Andreas
Danier, Duolikun
Vijayakumar, Sethu
Kouris, Alexandros
Mac Aodha, Oisin
Lu, Chris Xiaoxuan
Robotics
Computer Vision and Pattern Recognition
The adoption of pre-trained visual representations (PVRs), leveraging features from large-scale vision models, has become a popular paradigm for training visuomotor policies. However, these powerful representations can encode a broad range of task-irrelevant scene information, making the resulting trained policies vulnerable to out-of-domain visual changes and distractors. In this work we address visuomotor policy feature pooling as a solution to the observed lack of robustness in perturbed scenes. We achieve this via Attentive Feature Aggregation (AFA), a lightweight, trainable pooling mechanism that learns to naturally attend to task-relevant visual cues, ignoring even semantically rich scene distractors. Through extensive experiments in both simulation and the real world, we demonstrate that policies trained with AFA significantly outperform standard pooling approaches in the presence of visual perturbations, without requiring expensive dataset augmentation or fine-tuning of the PVR. Our findings show that ignoring extraneous visual information is a crucial step towards deploying robust and generalisable visuomotor policies. Project Page: tsagkas.github.io/afa
title Attentive Feature Aggregation or: How Policies Learn to Stop Worrying about Robustness and Attend to Task-Relevant Visual Cues
topic Robotics
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2511.10762