Incorporating Task-specific Concept Knowledge into Script Learning
This addresses script learning for AI systems by making it more practical with user context, though it appears incremental as it builds on existing script learning tasks.
The paper tackles the problem of Goal-Oriented Script Completion by introducing a more realistic setting that includes user context like preferences and history, and proposes a novel approach using concept prompting and script-oriented contrastive learning to improve performance on a WikiHow-based dataset.
In this paper, we present Tetris, a new task of Goal-Oriented Script Completion. Unlike previous work, it considers a more realistic and general setting, where the input includes not only the goal but also additional user context, including preferences and history. To address this problem, we propose a novel approach, which uses two techniques to improve performance: (1) concept prompting, and (2) script-oriented contrastive learning that addresses step repetition and hallucination problems. On our WikiHow-based dataset, we find that both methods improve performance. The dataset, repository, and models will be publicly available to facilitate further research on this new task.