Preface

摘要

There are many pressing questions that need to be answered in the field of designing and parsing meaning representations. First and foremost, does the field have an existential crisis in the face of the powerful combination of distributed representations and deep neural networks that recently led to many advances in the field or Natural Language Processing? In other words, will the distributed representations displace the "traditional" symbolic meaning representations that have been in development for many decades? If the answer is no, what are the areas where the symbolic meaning representation is most needed and has an advantage over distributed representations? There are currently also different schools of thoughts on how meaning should be represented. Can a consensus be forged on the key elements of meaning representation that are most needed given our current understanding of the needs in natural language applications? Regardless of the breadth of their use in natural language processing, how can meaning representations help us to advance theoretical linguistic research?