Recent Trends in Natural Language Understanding for Procedural Knowledge

In the last decade, there has been a surge of work contributing to commonsense reasoning and language understanding. Several datasets and knowledge representations for question answering, coreference resolution, and inference tasks have been discussed and built upon in the NLP community. However, these focus on declarative knowledge, or factual knowledge describing an entity or event. In contrast, a large body of research that has not been discussed, is reasoning and language understanding with procedural/how-to knowledge, or the knowledge used in completing a task or process (e.g., navigation instructions, recipes, repair guides, etc.). Several procedural knowledge bases, benchmark datasets, and information extraction methods, have enabled the various natural language understanding tasks with procedural knowledge (also known as procedural knowledge understanding). This paper seeks to provide an overview of the work in procedural knowledge understanding, and information extraction, acquisition, and representation with procedures, to promote discussion and provide a better understanding of procedural knowledge applications and future challenges.