Transferring Procedural Knowledge across Commonsense Tasks
This addresses the need for AI agents to understand everyday stories, though it appears incremental as it builds on existing supervised methods for story completion.
The paper tackles the problem of AI models lacking mechanisms to automatically track and explain procedures in unseen stories by developing LEAP, a framework that transfers procedural knowledge to novel narrative tasks, with experiments showing its labeler improves out-of-domain performance and provides explainability.
Stories about everyday situations are an essential part of human communication, motivating the need to develop AI agents that can reliably understand these stories. Despite the long list of supervised methods for story completion and procedural understanding, current AI has no mechanisms to automatically track and explain procedures in unseen stories. To bridge this gap, we study the ability of AI models to transfer procedural knowledge to novel narrative tasks in a transparent manner. We design LEAP: a comprehensive framework that integrates state-of-the-art modeling architectures, training regimes, and augmentation strategies based on both natural and synthetic stories. To address the lack of densely annotated training data, we devise a robust automatic labeler based on few-shot prompting to enhance the augmented data. Our experiments with in- and out-of-domain tasks reveal insights into the interplay of different architectures, training regimes, and augmentation strategies. LEAP's labeler has a clear positive impact on out-of-domain datasets, while the resulting dense annotation provides native explainability.