← Back to the feed
Robotics & manufacturingResearch findings

Cleanly-DINO: A Method for Learning Visual Task Spaces in Robotics

Original published 2026.11.16 02:11 KSTAdded to Blackdeck 2026.11.18 01:33 KST

What has changed?

Proposes a method to learn visual task spaces based on visual state changes, especially for cleaning tasks

Key points

  1. Cleaning tasks are defined by visual state changes, not object poses
  2. Cleanly-DINO learns visual task spaces to detect incomplete regions
  3. Enables cross-domain transfer and serves as observation encoder and reward signal in reinforcement learning

Connect it to your work

Effectively detects incomplete task regions by learning visual state progression

What remains unproven

This research note is based on an abstract; check the original for test conditions and application scope.

Originals used for this note

Links open the original publisher in a new tab.

  1. arxivCleanly-DINO: Efficiently Learning Visual Task Spaces for Robotics