Simulating Action Dynamics with Neural Process Networks
This work addresses the challenge of interpreting procedural text for AI systems, offering an incremental improvement over existing memory architectures.
The paper tackled the problem of understanding procedural language by anticipating causal effects of actions, introducing Neural Process Networks to simulate action dynamics and update entity states with learned operators, resulting in more accurate contextual information and interpretable representations.
Understanding procedural language requires anticipating the causal effects of actions, even when they are not explicitly stated. In this work, we introduce Neural Process Networks to understand procedural text through (neural) simulation of action dynamics. Our model complements existing memory architectures with dynamic entity tracking by explicitly modeling actions as state transformers. The model updates the states of the entities by executing learned action operators. Empirical results demonstrate that our proposed model can reason about the unstated causal effects of actions, allowing it to provide more accurate contextual information for understanding and generating procedural text, all while offering more interpretable internal representations than existing alternatives.