A major part of natural language processing now depends on the use of text data to build linguistic analyzers. We consider statistical, computational approaches to modeling linguistic structure. We seek to unify across many approaches and many kinds of linguistic structures. Assuming a basic understanding of natural language processing and/or machine learning, we seek to bridge the gap between the two fields. Approaches to decoding (i.e., carrying out linguistic structure prediction) and supervised and unsupervised learning of models that predict discrete structures as outputs are the focus. We also survey natural language processing problems to which these methods are being applied, and we address related topics in probabilistic inference, optimization, and experimental methodology. Table of Contents: Representations and Linguistic Data / Decoding: Making Predictions / Learning Structure from Annotated Data / Learning Structure from Incomplete Data / Beyond Decoding: Inference

A major part of natural language processing now depends on the use of text data to build linguistic analyzers. We consider statistical, computational approaches to modeling linguistic structure. We seek to unify across many approaches and many kinds of linguistic structures. Assuming a basic understanding of natural language processing and/or machine learning, we seek to bridge the gap between the two fields. Approaches to decoding (i.e., carrying out linguistic structure prediction) and supervised and unsupervised learning of models that predict discrete structures as outputs are the focus. We also survey natural language processing problems to which these methods are being applied, and we address related topics in probabilistic inference, optimization, and experimental methodology. Table of Contents: Representations and Linguistic Data / Decoding: Making Predictions / Learning Structure from Annotated Data / Learning Structure from Incomplete Data / Beyond Decoding: Inference

There is ample evidence that language users, including second-language (L2) users, can predict upcoming information during listening and reading. Yet it is still unclear when, how, and why language users engage in prediction, and what the relation is between prediction and learning. This volume presents a collection of current research, insights, and directions regarding the role of prediction in L2 processing and learning. The contributions in this volume specifically address how different (L1-based) theoretical models of prediction apply to or may be expanded to account for L2 processing, report new insights on factors (linguistic, cognitive, social) that modulate L2 users’ engagement in prediction, and discuss the functions that prediction may or may not serve in L2 processing and learning. Taken together, this volume illustrates various fruitful approaches to investigating and accounting for differences in predictive processing within and across individuals, as well as across populations.

Thomas Berg challenges context-free theories of linguistics; he is concerned with the way the term 'explanation' is typically used in the discipline. He argues that real explanations cannot emerge from a view which asserts the autonomy of language, but only from an approach which seeks to establish a connection between language and the contexts in which it is embedded. The author examines the psychological context in detail. He uses an interactiveactivation model of language processing to derive predictions about synchronic linguistic patterns, the course of linguistic change, and the structure of poetic rhymes. The majority of these predictions are borne out, leading the author to conclude that the structure of language is shaped by the properties of the mechanism which puts it to use, and that psycholinguistics thus qualifies as one likely approach from which to derive an explanation of linguistic structure.

An overview of recent work in the field of structured prediction, the building of predictive machine learning models for interrelated and dependent outputs. The goal of structured prediction is to build machine learning models that predict relational information that itself has structure, such as being composed of multiple interrelated parts. These models, which reflect prior knowledge, task-specific relations, and constraints, are used in fields including computer vision, speech recognition, natural language processing, and computational biology. They can carry out such tasks as predicting a natural language sentence, or segmenting an image into meaningful components. These models are expressive and powerful, but exact computation is often intractable. A broad research effort in recent years has aimed at designing structured prediction models and approximate inference and learning procedures that are computationally efficient. This volume offers an overview of this recent research in order to make the work accessible to a broader research community. The chapters, by leading researchers in the field, cover a range of topics, including research trends, the linear programming relaxation approach, innovations in probabilistic modeling, recent theoretical progress, and resource-aware learning. Contributors Jonas Behr, Yutian Chen, Fernando De La Torre, Justin Domke, Peter V. Gehler, Andrew E. Gelfand, Sébastien Giguère, Amir Globerson, Fred A. Hamprecht, Minh Hoai, Tommi Jaakkola, Jeremy Jancsary, Joseph Keshet, Marius Kloft, Vladimir Kolmogorov, Christoph H. Lampert, François Laviolette, Xinghua Lou, Mario Marchand, André F. T. Martins, Ofer Meshi, Sebastian Nowozin, George Papandreou, Daniel Průša, Gunnar Rätsch, Amélie Rolland, Bogdan Savchynskyy, Stefan Schmidt, Thomas Schoenemann, Gabriele Schweikert, Ben Taskar, Sinisa Todorovic, Max Welling, David Weiss, Thomáš Werner, Alan Yuille, Stanislav Živný

University of Massachusetts Occasional Papers in Linguistics, Vol. 38: Processing Linguistic Structure

This book constitutes the refereed proceedings of the 12th International Conference on Language and Automata Theory and Applications, LATA 2018, held in Ramat Gan, Israel, in April 2018.The 20 revised full papers presented together with 3 invited papers were carefully reviewed and selected from 58 submissions. The papers cover fields like algebraic language theory, algorithms for semi-structured data mining, algorithms on automata and words, automata and logic, automata for system analysis and programme verification, automata networks, automatic structures, codes, combinatorics on words, computational complexity, concurrency and Petri nets, data and image compression, descriptional complexity, foundations of finite state technology, foundations of XML, grammars (Chomsky hierarchy, contextual, unification, categorial, etc.), grammatical inference and algorithmic learning, graphs and graph transformation, language varieties and semigroups, language-based cryptography, mathematical and logical foundations of programming methodologies, parallel and regulated rewriting, parsing, patterns, power series, string processing algorithms, symbolic dynamics, term rewriting, transducers, trees, tree languages and tree automata, and weighted automata.

Natural language processing (NLP) went through a profound transformation in the mid-1980s when it shifted to make heavy use of corpora and data-driven techniques to analyze language. Since then, the use of statistical techniques in NLP has evolved in several ways. One such example of evolution took place in the late 1990s or early 2000s, when full-fledged Bayesian machinery was introduced to NLP. This Bayesian approach to NLP has come to accommodate for various shortcomings in the frequentist approach and to enrich it, especially in the unsupervised setting, where statistical learning is done without target prediction examples. We cover the methods and algorithms that are needed to fluently read Bayesian learning papers in NLP and to do research in the area. These methods and algorithms are partially borrowed from both machine learning and statistics and are partially developed "in-house" in NLP. We cover inference techniques such as Markov chain Monte Carlo sampling and variational inference, Bayesian estimation, and nonparametric modeling. We also cover fundamental concepts in Bayesian statistics such as prior distributions, conjugacy, and generative modeling. Finally, we cover some of the fundamental modeling techniques in NLP, such as grammar modeling and their use with Bayesian analysis.

A survey of computational methods for understanding, generating, and manipulating human language, which offers a synthesis of classical representations and algorithms with contemporary machine learning techniques. This textbook provides a technical perspective on natural language processing—methods for building computer software that understands, generates, and manipulates human language. It emphasizes contemporary data-driven approaches, focusing on techniques from supervised and unsupervised machine learning. The first section establishes a foundation in machine learning by building a set of tools that will be used throughout the book and applying them to word-based textual analysis. The second section introduces structured representations of language, including sequences, trees, and graphs. The third section explores different approaches to the representation and analysis of linguistic meaning, ranging from formal logic to neural word embeddings. The final section offers chapter-length treatments of three transformative applications of natural language processing: information extraction, machine translation, and text generation. End-of-chapter exercises include both paper-and-pencil analysis and software implementation. The text synthesizes and distills a broad and diverse research literature, linking contemporary machine learning techniques with the field's linguistic and computational foundations. It is suitable for use in advanced undergraduate and graduate-level courses and as a reference for software engineers and data scientists. Readers should have a background in computer programming and college-level mathematics. After mastering the material presented, students will have the technical skill to build and analyze novel natural language processing systems and to understand the latest research in the field.

This book aims to contribute to a growing interest amongst psycholinguists and morphologists in the mechanisms of meaning predictability. It presents a brand-new model of the meaning-prediction of novel, context-free naming units, relating the wordformation and wordinterpretation processes. Unlike previous studies, mostly focussed on N+N compounds, the scope of this book is much wider. It not only covers all types of complex words, but also discusses a whole range of predictability-boosting and -reducing conditions. Two measures are introduced, the Predictability Rate and the Objectified Predictability Rate, in order to compare the strength of predictable readings both within a word and relative to the most predictable readings of other coinages. Four extensive experiments indicate inter alia the equal predicting capacity of native and non-native speakers, the close interconnection between linguistic and extra-linguistic factors, the important role of prototypical semes, and the usual dominance of a single central reading.