LISA: Learning Interpretable Skill Abstractions from Language
This work addresses generalization issues in language-conditioned sequential decision-making for navigation and robotic manipulation, offering an incremental improvement with interpretable skill learning.
The paper tackles the problem of learning policies that effectively use language instructions in multi-task environments by proposing LISA, a hierarchical imitation learning framework that learns interpretable skills from language-conditioned demonstrations. The result shows that LISA outperforms a Decision Transformer baseline in low-data regimes and can compose skills to solve tasks with unseen instructions.
Learning policies that effectively utilize language instructions in complex, multi-task environments is an important problem in sequential decision-making. While it is possible to condition on the entire language instruction directly, such an approach could suffer from generalization issues. In our work, we propose \emph{Learning Interpretable Skill Abstractions (LISA)}, a hierarchical imitation learning framework that can learn diverse, interpretable primitive behaviors or skills from language-conditioned demonstrations to better generalize to unseen instructions. LISA uses vector quantization to learn discrete skill codes that are highly correlated with language instructions and the behavior of the learned policy. In navigation and robotic manipulation environments, LISA outperforms a strong non-hierarchical Decision Transformer baseline in the low data regime and is able to compose learned skills to solve tasks containing unseen long-range instructions. Our method demonstrates a more natural way to condition on language in sequential decision-making problems and achieve interpretable and controllable behavior with the learned skills.