GC-TTT adapts goal-conditioned policies at test time by fine-tuning on self-supervised selected goal-related offline data, yielding performance gains in loco-navigation and manipulation tasks.
D.1 Hyperparameters GC-TTT introduces some additional hyperparameters
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Test-time Offline Reinforcement Learning on Goal-related Experience
GC-TTT adapts goal-conditioned policies at test time by fine-tuning on self-supervised selected goal-related offline data, yielding performance gains in loco-navigation and manipulation tasks.