CAREL improves instruction-following RL sample efficiency by aligning observation sequences with instruction tokens via an X-CLIP style contrastive loss and masking completed subtasks.
Modular multitask reinforcement learning with policy sketches
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
CAREL: Instruction-guided reinforcement learning with cross-modal auxiliary objectives
CAREL improves instruction-following RL sample efficiency by aligning observation sequences with instruction tokens via an X-CLIP style contrastive loss and masking completed subtasks.