DIFF-IL combines per-frame domain-invariant feature extraction with frame-wise time labeling to improve cross-domain imitation learning from images, beating prior methods on 14 tasks.
Robust imitation learning for mobile manipulator focusing on task-related viewpoints and regions
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Domain-Invariant Per-Frame Feature Extraction for Cross-Domain Imitation Learning with Visual Observations
DIFF-IL combines per-frame domain-invariant feature extraction with frame-wise time labeling to improve cross-domain imitation learning from images, beating prior methods on 14 tasks.