In recent years there has been a huge amount of interest and work from the research community on the use of deep neural architectures for Natural Language Processing (NLP). But despite the fact that automatic representation learning, as opposed to manual feature engineering, has become the de facto standard methodological framework, linguistic knowledge - encoded informally in the form of human expertise and intuitions about the language, but also formally in large symbolic linguistic resources - has not become obsolete. It is still an invaluable source of knowledge required by most modern NLP technologies to reach peak performance. For instance, for high-end, complex NLP tasks the amount of available training samples could still be insufficient, motivating the need for efficient task specific learning biases. For this purpose, a wealth of linguistic resources of various kinds - which encode large amounts of facts about languages and general-purpose background knowledge - could be used. They constrain and bias the learning process of deep neural architectures.
This special issue aims to collect state-of-the-art contributions to the development and use of linguistic and background knowledge for neural architectures in NLP. These include but are not limited to: a) task-specific objective functions that are informed on the basis of linguistic knowledge; b) using linguistic resources like lexical knowledge bases and multilingual dictionaries to generate training data for neural architectures, as well as specialize and improve text representations; c) leveraging syntactic and semantic information in sequence-to-sequence and other neural models. We are particularly interested in contributions showing the benefits of using linguistic and deep neural techniques synergistically, as well as how these complement each other to advance the state of the art in our field. One instance of such technologies are methods that embed into vector representations of words information about symbolic linguistic structures of linguistic graphs and knowledge bases.
Topics of interest include, but are not limited to:
- Neural models that rely on existing repositories of linguistic knowledge, such as lexical semantic resources or linguistically annotated corpora. - New ways and principles for integration of linguistic knowledge into the modern neural architectures. - Specialization of word embeddings to specific types of semantic relations, e.g., hypernymy, synonymy, antonymy, etc. - Learning word and graph embeddings based on large knowledge bases and other graphs of linguistic information. - Learning word sense embeddings based on existing lexical semantic resources. - Applications of hybrid neural models in NLP tasks, such as question answering, text summarization, machine translation, and natural language understanding. - Neural methods for creating, improving and enriching linguistic resources.
The deadline for submissions is the 1st of November, 2018. Please submit your article using NLE editorial system at http://mc.manuscriptcentral.com/nle and select “SI: DL and Linguistics” during the submission process.
- Submission deadline: November 1, 2018 - First-round author notification: February 1, 2019 - Revised versions due: May 1, 2019 - Second-round author notification: June 1, 2019 - Final versions due: July 15, 2019 - Special issue publication: fall/winter 2019
Best - Simone (with co-editors Ivan and Alex)