Language-Conditioned Semantic Search-Based Policy for Robotic Manipulation Tasks
This addresses the problem of limited generalization in robotic manipulation for tasks typically handled by imitation or reinforcement learning, though it appears incremental as it builds on existing search-based and language-conditioned approaches.
The paper tackles the challenge of generalizing robotic manipulation policies from few demonstrations by proposing a language-conditioned semantic search method that retrieves actions from similar trajectories in a dataset. It outperforms baselines on the CALVIN benchmark and shows strong zero-shot adaptation capabilities.
Reinforcement learning and Imitation Learning approaches utilize policy learning strategies that are difficult to generalize well with just a few examples of a task. In this work, we propose a language-conditioned semantic search-based method to produce an online search-based policy from the available demonstration dataset of state-action trajectories. Here we directly acquire actions from the most similar manipulation trajectories found in the dataset. Our approach surpasses the performance of the baselines on the CALVIN benchmark and exhibits strong zero-shot adaptation capabilities. This holds great potential for expanding the use of our online search-based policy approach to tasks typically addressed by Imitation Learning or Reinforcement Learning-based policies.