Accompanying continued industrial production and sales of artificial intelligence and expert systems is the risk that difficult and resistant theoretical problems and issues will be ignored. The participants at the Third Tinlap Workshop, whose contributions are contained in Theoretical Issues in Natural Language Processing, remove that risk. They discuss and promote theoretical research on natural language processing, examinations of solutions to current problems, development of new theories, and representations of published literature on the subject. Discussions among these theoreticians in artificial intelligence, logic, psychology, philosophy, and linguistics draw a comprehensive, up-to-date picture of the natural language processing field.

Three years ago (June 6-8, 1975) a highly successful interdisciplinary workshop called TINLAP (Theoretical Issues in Natural Language Processing) was held at MIT with the purpose of "bringing together researchers and students from computational linguistics, psychology, linguistics and artificial intelligence to provide a forum at which people with different interests in, and consequently different emphases on, the problems of natural language understanding, could learn of the models developed and difficult issues faced by people working on other aspects of understanding." The workshop was organized by Roger Schank of Yale and Bonnie For TINLAP-2, I selected six topics which I felt were especially interesting from among those suggested by persons who had asked for another TINLAP meeting, and generated lists of questions to be considered by authors for four of the sessions; Session 2 questions were generated oy Bonnie Lynn Webber, and those in Session 5 are from Aravind Joshi. The lists of questions appear at the end of this volume. Authors were encouraged to summarize or criticize research, to speculate on language understanding mechanisms or future research directions, or to report recent work of their own.

Multilingual Natural Language Processing Applications is the first comprehensive single-source guide to building robust and accurate multilingual NLP systems. Edited by two leading experts, it integrates cutting-edge advances with practical solutions drawn from extensive field experience. Part I introduces the core concepts and theoretical foundations of modern multilingual natural language processing, presenting today’s best practices for understanding word and document structure, analyzing syntax, modeling language, recognizing entailment, and detecting redundancy. Part II thoroughly addresses the practical considerations associated with building real-world applications, including information extraction, machine translation, information retrieval/search, summarization, question answering, distillation, processing pipelines, and more. This book contains important new contributions from leading researchers at IBM, Google, Microsoft, Thomson Reuters, BBN, CMU, University of Edinburgh, University of Washington, University of North Texas, and others. Coverage includes Core NLP problems, and today’s best algorithms for attacking them Processing the diverse morphologies present in the world’s languages Uncovering syntactical structure, parsing semantics, using semantic role labeling, and scoring grammaticality Recognizing inferences, subjectivity, and opinion polarity Managing key algorithmic and design tradeoffs in real-world applications Extracting information via mention detection, coreference resolution, and events Building large-scale systems for machine translation, information retrieval, and summarization Answering complex questions through distillation and other advanced techniques Creating dialog systems that leverage advances in speech recognition, synthesis, and dialog management Constructing common infrastructure for multiple multilingual text processing applications This book will be invaluable for all engineers, software developers, researchers, and graduate students who want to process large quantities of text in multiple languages, in any environment: government, corporate, or academic.

Embeddings have undoubtedly been one of the most influential research areas in Natural Language Processing (NLP). Encoding information into a low-dimensional vector representation, which is easily integrable in modern machine learning models, has played a central role in the development of NLP. Embedding techniques initially focused on words, but the attention soon started to shift to other forms: from graph structures, such as knowledge bases, to other types of textual content, such as sentences and documents. This book provides a high-level synthesis of the main embedding techniques in NLP, in the broad sense. The book starts by explaining conventional word vector space models and word embeddings (e.g., Word2Vec and GloVe) and then moves to other types of embeddings, such as word sense, sentence and document, and graph embeddings. The book also provides an overview of recent developments in contextualized representations (e.g., ELMo and BERT) and explains their potential in NLP. Throughout the book, the reader can find both essential information for understanding a certain topic from scratch and a broad overview of the most successful techniques developed in the literature.