Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This study investigates the effectiveness of segmented action annotations as an interface for fine-tuning variable-length action (VLA) policies in robotic manipulation tasks. By comparing the performance of different representations of stage information—current-stage text and normalized ordinal stage indices—under both direct and continuation fine-tuning, the authors find that explicit stage information does not universally enhance policy performance. Notably, while the Ordinal Stage-State representation outperforms the alternatives in continuation fine-tuning, the overall results suggest that the benefits of stage information are context-dependent and vary with training methods.
Explicit stage information in fine-tuning doesn't guarantee improved policy performance, revealing a nuanced relationship between interface design and training strategy.
One high-level instruction in long-horizon manipulation can cover several action stages. We use segmented action annotations as an intermediate representation between the full-task instruction and VLA action chunks. A progress module tracks the active stage, while the action policy receives stage information either as current-stage text or as a normalized ordinal stage index in robot state. We compare these interfaces with GR00T N1.6 on LIBERO-10 under direct fine-tuning and continuation fine-tuning from a full-task instruction baseline. Under direct fine-tuning, full-task instruction, current-stage text, and Ordinal Stage-State achieve mean success rates of 57.45%, 50.24%, and 54.36%, respectively, showing that explicit stage information does not automatically improve the policy. Under continuation, the corresponding means are 49.07%, 50.00%, and 53.75%, with Ordinal Stage-State exceeding both alternatives in all three paired runs. The observed benefit differs across interface representations and training arrangements.