Dissertations in Artificial Intelligence (IOS Press) – serie

Word embeddings are a form of distributional semantics increasingly popular for investigating lexical semantic change. However, typical training algorithms are probabilistic, limiting their reliability and the reproducibility of studies.Johannes Hellrich investigated this problem both empirically and theoretically and found some variants of SVD-based algorithms to be unaffected. Furthermore, he created the JeSemE website to make word embedding based diachronic research more accessible. It provides information on changes in word denotation and emotional connotation in five diachronic corpora.Finally, the author conducted two case studies on the applicability of these methods by investigating the historical understanding of electricity as well as words connected to Romanticism. They showed the high potential of distributional semantics for further applications in the digital humanities.

A common paradigm in distance-based learning is to embed the instance space into a feature space equipped with a metric and define the dissimilarity between instances by the distance of their images in the feature space. Frequent connected subgraphs are sometimes used to define such feature spaces if the instances are graphs, but identifying the set of frequent connected subgraphs and subsequently computing embeddings for graph instances is computationally intractable. As a result, existing frequent subgraph mining algorithms either restrict the structural complexity of the instance graphs or require exponential delay between the output of subsequent patterns, meaning that distance-based learners lack an efficient way to operate on arbitrary graph data.This book presents a mining system that gives up the demand on the completeness of the pattern set, and instead guarantees a polynomial delay between subsequent patterns. To complement this, efficient methods devised to compute the embedding of arbitrary graphs into the Hamming space spanned by the pattern set are described. As a result, a system is proposed that allows the efficient application of distance-based learning methods to arbitrary graph databases. In addition to an introduction and conclusion, the book is divided into chapters covering: preliminaries; related work; probabilistic frequent subtrees; boosted probabilistic frequent subtrees; and fast computation, with a further two chapters on Hamiltonian path for cactus graphs and Poisson binomial distribution.

A core problem in Artificial Intelligence is the modeling of human reasoning. Classic-logical approaches are too rigid for this task, as deductive inference yielding logically correct results is not appropriate in situations where conclusions must be drawn based on the incomplete or uncertain knowledge present in virtually all real world scenarios.Since there are no mathematically precise and generally accepted definitions for the notions of plausible or rational, the question of what a knowledge base consisting of uncertain rules entails has long been an issue in the area of knowledge representation and reasoning. Different nonmonotonic logics and various semantic frameworks and axiom systems have been developed to address this question.The main theme of this book, Knowledge Representation and Inductive Reasoning using Conditional Logic and Sets of Ranking Functions, is inductive reasoning from conditional knowledge bases. Using ordinal conditional functions as ranking models for conditional knowledge bases, the author studies inferences induced by individual ranking models as well as by sets of ranking models. He elaborates in detail the interrelationships among the resulting inference relations and shows their formal properties with respect to established inference axioms. Based on the introduction of a novel classification scheme for conditionals, he also addresses the question of how to realize and implement the entailment relations obtained.In this work, “Steven Kutsch convincingly presents his ideas, provides illustrating examples for them, rigorously defines the introduced concepts, formally proves all technical results, and fully implements every newly introduced inference method in an advanced Java library (…). He significantly advances the state of the art in this field.” – Prof. Dr. Christoph Beierle of the FernUniversität in Hagen

The last few decades have seen impressive improvements in several areas of Natural Language Processing. Nevertheless, getting a computer to make sense of the discourse of utterances in a text remains challenging. Several different theories which aim to describe and analyze the coherent structure of a well-written text exist, but with varying degrees of applicability and feasibility for practical use. This book is about shallow discourse parsing, following the paradigm of the Penn Discourse TreeBank, a corpus containing over 1 million words annotated for discourse relations. When it comes to discourse processing, any language other than English must be considered a low-resource language. This book relates to discourse parsing for German. The limited availability of annotated data for German means that the potential of modern, deep-learning-based methods relying on such data is also limited. This book explores to what extent machine-learning and more recent deep-learning-based methods can be combined with traditional, linguistic feature engineering to improve performance for the discourse parsing task.The end-to-end shallow discourse parser for German developed for the purpose of this book is open-source and available online. Work has also been carried out on several connective lexicons in different languages. Strategies are discussed for creating or further developing such lexicons for a given language, as are suggestions on how to further increase their usefulness for shallow discourse parsing.The book will be of interest to all whose work involves Natural Language Processing, particularly in languages other than English.

One of the core problems in artificial intelligence is the modelling of human reasoning and intelligent behaviour. The representation of knowledge, and reasoning about it, are of crucial importance in achieving this. This book, Semantics of Belief Change Operators for Intelligent Agents: Iteration, Postulates, and Realizability, addresses a number of significant research questions in belief change theory from a semantic point of view; in particular, the connection between different types of belief changes and plausibility relations over possible worlds is investigated. This connection is characterized for revision over general classical logics, showing which relations are capturing AGM revision. In addition, those classical logics for which the correspondence between AGM revision and total preorders holds are precisely characterized. AGM revision in the Darwiche-Pearl framework for belief change over arbitrary sets of epistemic states is considered, demonstrating, especially, that for some sets of epistemic states, no AGM revision operator exists. A characterization of those sets of epistemic states for which AGM revision operators exist is presented. The expressive class of dynamic limited revision operators is introduced to provide revision operators for more sets of epistemic states. Specifications for the acceptance behaviour of various belief-change operators are examined, and those realizable by dynamic-limited revision operators are described. The iteration of AGM contraction in the Darwiche-Pearl framework is explored in detail, several known and novel iteration postulates for contraction are identified, and the relationships among these various postulates are determined.With a convincing presentation of ideas, the book refines and advances existing proposals of belief change, develops novel concepts and approaches, rigorously defines the concepts introduced, and formally proves all technical claims, propositions and theorems, significantly advancing the state-of-the-art in this field.

Although both deal with narratives, the two disciplines of Narrative Theory (NT) and Computational Story Composition (CSC) rarely exchange insights and ideas or engage in collaborative research. The former has its roots in the humanities, and attempts to analyze literary texts to derive an understanding of the concept of narrative. The latter is in the domain of Artificial Intelligence, and investigates the autonomous composition of fictional narratives in a way that could be deemed creative. The two disciplines employ different research methodologies at contradistinct levels of abstraction, making simultaneous research difficult, while a close exchange between the two disciplines would undoubtedly be desirable, not least because of the complementary approach to their object of study.This book, From Narratology to Computational Story Composition and Back, describes an exploratory study in generative modeling, a research methodology proposed to address the methodological differences between the two disciplines and allow for simultaneous NT and CSC research. It demonstrates how implementing narratological theories as computational, generative models can lead to insights for NT, and how grounding computational representations of narrative in NT can help CSC systems to take over creative responsibilities. It is the interplay of these two strands that underscores the feasibility and utility of generative modeling. The book is divided into 6 chapters: an introduction, followed by chapters on plot, fictional characters, plot quality estimation, and computational creativity, wrapped up by a conclusion.The book will be of interest to all those working in the fields of narrative theory and computational creativity.

Although both deal with narratives, the two disciplines of Narrative Theory (NT) and Computational Story Composition (CSC) rarely exchange insights and ideas or engage in collaborative research. The former has its roots in the humanities, and attempts to analyze literary texts to derive an understanding of the concept of narrative. The latter is in the domain of Artificial Intelligence, and investigates the autonomous composition of fictional narratives in a way that could be deemed creative. The two disciplines employ different research methodologies at contradistinct levels of abstraction, making simultaneous research difficult, while a close exchange between the two disciplines would undoubtedly be desirable, not least because of the complementary approach to their object of study.This book, From Narratology to Computational Story Composition and Back, describes an exploratory study in generative modeling, a research methodology proposed to address the methodological differences between the two disciplines and allow for simultaneous NT and CSC research. It demonstrates how implementing narratological theories as computational, generative models can lead to insights for NT, and how grounding computational representations of narrative in NT can help CSC systems to take over creative responsibilities. It is the interplay of these two strands that underscores the feasibility and utility of generative modeling. The book is divided into 6 chapters: an introduction, followed by chapters on plot, fictional characters, plot quality estimation, and computational creativity, wrapped up by a conclusion.The book will be of interest to all those working in the fields of narrative theory and computational creativity.

Traditional workflow management systems support the fulfillment of business tasks by providing guidance along a predefined workflow model. Due to the shift from mass production to customization, flexibility has become important in recent decades, but the various approaches to workflow flexibility either require extensive knowledge acquisition and modeling, or active intervention during execution. Pursuing flexibility by deviation compensates for these disadvantages by allowing alternative paths of execution at run time without requiring adaptation to the workflow model.This work, Flexible Workflows: A Constraint- and Case-Based Approach, proposes a novel approach to flexibility by deviation, the aim being to provide support during the execution of a workflow by suggesting items based on predefined strategies or experiential knowledge, even in case of deviations. The concepts combine two familiar methods from the field of AI - constraint satisfaction problem solving, and process-oriented case-based reasoning. The combined model increases the capacity for flexibility. The experimental evaluation of the approach consisted of a simulation involving several types of participant in the domain of deficiency management in construction. The book contains 7 chapters covering foundations; domains and potentials; prerequisites; constraint based workflow engine; case based deviation management; prototype; and evaluation, together with an introduction, a conclusion and 3 appendices.Demonstrating high utility values and the promise of wide applicability in practice, as well as the potential for an investigation into the transfer of the approach to other domains, the book will be of interest to all those whose work involves workflow management systems.