Showing papers on "Natural language understanding published in 2008"

Spoken language understanding

Abstract: Semantics deals with the organization of meanings and the relations between sensory signs or symbols and what they denote or mean. Computational semantics performs a conceptualization of the world using computational processes for composing a meaning representation structure from available signs and their features present, for example, in words and sentences. Spoken language understanding (SLU) is the interpretation of signs conveyed by a speech signal. SLU and natural language understanding (NLU) share the goal of obtaining a conceptual representation of natural language sentences. Specific to SLU is the fact that signs to be used for interpretation are coded into signals along with other information such as speaker identity. Furthermore, spoken sentences often do not follow the grammar of a language; they exhibit self-corrections, hesitations, repetitions, and other irregular phenomena. SLU systems contain an automatic speech recognition (ASR) component and must be robust to noise due to the spontaneous nature of spoken language and the errors introduced by ASR. Moreover, ASR components output a stream of words with no structure information like punctuation and sentence boundaries. Therefore, SLU systems cannot rely on such markers and must perform text segmentation and understanding at the same time.

Toward Human Level Machine Intelligence - Is It Achievable? The Need for a Paradigm Shift

Abstract: Officially, AI was born in 1956. Since then, very impressive progress has been made in many areas - but not in the realm of human level machine intelligence. During much of its early history, AI "was rife "with exaggerated expectations. A headline in an article published in the late forties of last century was headlined, "Electric brain capable of translating foreign languages is being built". Today, more than half a century later, we do have translation software, but nothing that can approach the quality of human translation. Clearly, achievement of human level machine intelligence is a challenge that is hard to meet. A prerequisite to achievement of human level machine intelligence is mechanization of these capabilities and, in particular, mechanization of natural language understanding. To make significant progress toward achievement of human level machine intelligence, a paradigm shift is needed. More specifically, what is needed is an addition to the armamentarium of AI of two methodologies: (a) a nontraditional methodology of computing with words (CW) or more generally, NL-Computation; and (b) a countertraditional methodology "which involves a progression from computing with numbers to computing with words. The centerpiece of these methodologies is the concept of precisiation of meaning. Addition of these methodologies to AI would be an important step toward the achievement of human level machine intelligence and its applications in decision-making, pattern recognition, analysis of evidence, diagnosis, and assessment of causality. Such applications have a position of centrality in our infocentric society.

Abduction in Natural Language Understanding

Abstract: it is the street that is in Chicago. Therefore, a large part of the study of language should be an investigation of the question of how we use our knowledge of the world to understand discourse. This question has been examined primarily by researchers in the eld of arti cial intelligence (AI), in part because they have been interested in linking language with actual behavior in speci c situations, which has led them to an attempt to represent and reason about fairly complex world knowledge. In this chapter I describe how a particular kind of reasoning, called abduction, provides a framework for addressing a broad range of problems that are posed in discourse and that require world knowledge in their solutions. I rst defend rst-order logic as a mode of representation for the information conveyed by sentences and the knowledge we bring to the discourses we interpret, but with one caveat: Reasoning must be defeasible. I discuss several ways that defeasible inference has been formalized in AI, and introduce abduction as one of those methods. Then in successive sections I show

Situation and context in data fusion and natural language understanding

Abstract: This paper defines and relates several important concepts in data fusion and natural language understanding: situation, relation, relationship and context. In data fusion - as in other problem-solving applications - contextual reasoning involves inferring desired information (ldquoproblem variablesrdquo) on the basis of other available information (ldquocontext variablesrdquo). Relevant contexts are often not self-evident, but must be discovered or selected as a means to problems-solving. Therefore, context exploitation involves an integration of data fusion with planning and control functions. These concepts can be generalized to apply in very diverse context exploitation applications, to include natural language understanding; which similarly involves data alignment, association and estimation of speaker/ authorspsila intended meanings and references. Discovering and selecting useful context variables is an abductive data fusion/ management problem that can be characterized in a utility/ uncertainty framework.

Toward human level machine intelligence - is it achievable?

Abstract: Achievement of human level machine intelligence has long been one of the basic objectives of AI. Officially, AI was born in 1956. Since then, very impressive progress has been made in many areas - but not in the realm of human level machine intelligence. Anyone who has been forced to use a dumb automated customer service system will readily agree. The Turing test lies far beyond. Today, no machine can pass the Turing test and none is likely to do so in the foreseeable future.To make progress toward achievement of human level machine intelligence, AI must add to its armamentarium concepts and techniques drawn from other methodologies, especially evolutionary computing, neurocomputing and fuzzy logic. A key contribution of fuzzy logic is the machinery of Computing with Words (CW) and, more generally, NL-Computation. This machinery opens the door to mechanization of natural language understanding and computation with information described in natural language. Addition of this machinery to the armamentarium of AI would be an important step toward the achievement of human level machine intelligence and its applications in decision making, pattern recognition, analysis of evidence, diagnosis and assessment of causality. Such applications have a position of centrality in our info-centric society.

Spoken Dialogue Systems

Abstract: Spoken dialogue systems are a new breed of interfaces that enable humans to communicate with machines naturally and efficiently using a conversational paradigm. Such a system makes use of many human language technology (HLT) components, including speech recognition and synthesis, natural language understanding and generation, discourse modeling, and dialogue management. In this contribution, we introduce the nature of these interfaces, describe the underlying HLTs on which they are based, and discuss some of the development issues. After providing a historical perspective, we outline some new research directions.

Full text retrieval system based on natural language

Abstract: The invention discloses a full text retrieval system based on natural language understanding, comprising: a database server, an information receiving judging module, a natural language processing module, a retrieving module, an indexing module, an index database and a result set processing module. The system of the invention provides two resolution strategies, that is, word classification static with semantic analysis associated with automatic segmentation and expanding inquired word static according to Hownet rule for low intelligence situation of current search engine. The deployed system converts information retrieval from current key word-based layer to knowledge (or concept)-based layer; the invention is capable of using techniques such as word classification, synonym, concept search, phrase identification, etc. with understanding and processing ability to knowledge. The search engine is provided with intelligence and humanization of information service. The user is allowed using natural language for information retrieval. The invention is capable of adding user selection behavior in interactive operation mode, so as to provide more convenient, more precise search service.

Collecting Semantics in the Wild: The Story Workbench.

Abstract: Analogical reasoning is crucial to robust and flexible highlevel cognition. However, progress on computational models of analogy has been impeded by our inability to quickly and accurately collect large numbers (100+) of semantically annotated texts. The Story Workbench is a tool that facilitates such annotation by using natural language processing techniques to make a guess at the annotation, followed by approval, correction, and elaboration of that guess by a human annotator. Central to this approach is the use of a sophisticated graphical user interface that can guide even an untrained annotator through the annotation process. I describe five desiderata that govern the design of the Story Workbench, and demonstrate how each principle was fulfilled in the current implementation. The Story Workbench enables numerous experiments that previously were prohibitively laborious, of which I describe three currently underway in my lab. Analogical reasoning underlies many important cognitive processes, including learning, categorization, planning, and natural language understanding (Gentner, Holyoak, and Kokinov 2001). It is crucial to robust and flexible highlevel cognition. Despite great strides early in the computational understanding of analogical reasoning (Gick and Holyoak 1980; Winston 1980; Gentner 1983; Falkenhainer, Forbus, and Gentner 1989; Forbus, Gentner, and Law 1994), recent progress has been slow. Most computational models of analogy require semantic knowledge as input, supplied as semantically annotated texts. Historically, as the models became more complex, vetting them required ever larger sets of annotations, of greater detail and complexity. It is the assembly of these sets that has become a major bottleneck to progress. The sets should contain hundreds of annotations of sufficient richness, must have high interannotator agreement, and need to be collected quickly and without prohibitive expense. Manual assembly of such sets is costly, time-consuming, and error-prone. Automatic annotation systems also provide no relief: they lack coverage, are often imprecise or inaccurate, and are in general unable to provide the full scope of annotations required. This datacollection bottleneck has seriously impaired progress in the computational understanding of analogical reasoning, and a Copyright c © 2008, Association for the Advancement of Artificial Intelligence (www.aaai.org). All rights reserved. solution is needed if progress is to resume at a reasonable pace.

Virtual pet chatting system, method and virtual pet questioning and answering server

Abstract: A virtual pet chatting system includes a virtual pet client unit, a virtual pet data maintaining unit as well as a questioning and answering unit. A virtual pet chatting method includes: sending, by a first virtual pet, a natural language question to a second virtual pet; and generating, by the second virtual pet, a natural language response sentence according to the natural language question, after understanding the natural language and performing reasoning taking into account attributes of a virtual pet. A virtual pet questioning and answering server includes a natural language understanding module and a response sentence generating module.

A Common Ground for Virtual Humans: Using an Ontology in a Natural Language Oriented Virtual Human Architecture

Abstract: When dealing with large, distributed systems that use state-of-the-art components, individual components are usually developed in parallel. As development continues, the decoupling invariably leads to a mismatch between how these components internally represent concepts and how they communicate these representations to other components: representations can get out of synch, contain localized errors, or become manageable only by a small group of experts for each module. In this paper, we describe the use of an ontology as part of a complex distributed virtual human architecture in order to enable better communication between modules while improving the overall flexibility needed to change or extend the system. We focus on the natural language understanding capabilities of this architecture and the relationship between language and concepts within the entire system in general and the ontology in particular.

A Tool for Training Primary Health Care Medical Students: The Virtual Simulated Patient

Abstract: In this paper we present an embodied conversational agent (ECA) that simulates a real human patient presenting several symptoms for training medical students in the field of primary health care. Students interview the ECA to diagnose his diseases as a doctor would do in a real situation. This virtual patient can communicate using natural language and express different moods that depend on the diseases he suffers from and the studentpsilas behavior. The ECApsilas behavior is done by means of the coordination of several modules devoted to different tasks: natural language understanding, dialogue management, emotional state control and natural language generation. An ontology that gathers the domain knowledge of the agent specifies a semantic language that the modules use to communicate themselves. The system has two kinds of advantages. First, it offers a way to put in practice the theoretical knowledge acquired by the students in their degree courses and to improve the diagnostic and communicative skills decreasing the learning curve of these abilities, setting them for an interview process like the one they are going to face in their real job. Second, the system is ready to use at any time without needing special or expensive equipment, only a standard PC.

Historical Perspective of the Field of ASR/NLU

Abstract: The quest for a machine that can recognize and understand speech, from any speaker, and in any environment has been the holy grail of speech recognition research for more than 70 years. Although we have made great progress in understanding how speech is produced and analyzed, and although we have made enough advances to build and deploy in the field a number of viable speech recognition systems, we still remain far from the ultimate goal of a machine that communicates naturally with any human being. It is the goal of this section to document the history of research in speech recognition and natural language understanding, and to point out areas where great progress has been made, along with the challenges that remain to be solved in the future.

Utilizing Natural Language for One-Shot Task Learning

Abstract: Learning tasks from a single demonstration presents a significant challenge because the observed sequence is specific to the current situation and is inherently an incomplete representation of the procedure. Observation-based machine-learning techniques are not effective without multiple examples. However, when a demonstration is accompanied by natural language explanation, the language provides a rich source of information about the relationships between the steps in the procedure and the decision-making processes that led to them. In this article, we present a one-shot task learning system built on TRIPS, a dialogue-based collaborative problem solving system, and show how natural language understanding can be used for effective one-shot task learning.

Extraction of Socio-semantic Data from Chat Conversations in Collaborative Learning Communities

Abstract: Online collaboration among communities of practice using text-based tools, such as instant messaging, forums and web logs (blogs), has become very popular in the last years, but it is difficult to automatically analyze all their content due to the problems of natural language understanding software. However, useful socio-semantic data can be retrieved from a chat conversation using ontology-based text mining techniques. In this paper, a novel approach for detecting several kinds of semantic data from a chat conversation is presented. This method uses a combination of a dialogistic, socio-cultural perspective and of classical knowledge-based text processing methods. Lexical and domain ontologies are used. A tool has been developed for the discovery of the most important topics and of the contribution of each participant in the conversation. The system also discovers new, implicit references among the utterances of the chat in order to offer a multi-voiced representation of the conversation. The application offers a panel for visualizing the threading of the subjects in the chat and the contributions function. The system was experimented on chat sessions of small groups of students participating in courses on Human-Computer Interaction and Natural Language Processing in "Politehnica" University of Bucharest, Romania.

Information extraction tools and methods for understanding dialogue in a companion

Abstract: This paper discusses how Information Extraction is used to understand and manage Dialogue in the EU-funded Companions project. This will be discussed with respect to the Senior Companion, one of two applications under development in the EU-funded Companions project. Over the last few years, research in human-computer dialogue systems has increased and much attention has focused on applying learning methods to improving a key part of any dialogue system, namely the dialogue manager. Since the dialogue manager in all dialogue systems relies heavily on the quality of the semantic interpretation of the user’s utterance, our research in the Companions project, focuses on how to improve the semantic interpretation and combine it with knowledge from the Knowledge Base to increase the performance of the Dialogue Manager. Traditionally the semantic interpretation of a user utterance is handled by a natural language understanding module which embodies a variety of natural language processing techniques, from sentence splitting, to full parsing. In this paper we discuss the use of a variety of NLU processes and in particular Information Extraction as a key part of the NLU module in order to improve performance of the dialogue manager and hence the overall dialogue system.

Interface Model Elicitation from Textual Scenarios

Abstract: During the stage of system requirements gathering, model elicitation is aimed at identifying in textual scenarios model elements that are relevant for building a first version of models that will be further exploited in a model-driven engineering method. When multiple elements should be identified from multiple interrelated conceptual models, the complexity increases. Three method levels are successively examined to conduct model elicitation from textual scenarios for the purpose of conducting model-driven engineering of user interfaces: manual classification, dictionary-based classification, and nearly natural language understanding based on semantic tagging and chunk extraction. A model elicitation tool implementing these three levels is described and exemplified on a real-world case study for designing user interfaces to workflow information systems. The model elicitation process discussed in the case study involves several models: user, task, domain, organization, resources, and job.

Application of Bayesian Framework In Natural Language Understanding

Abstract: A natural language understanding (NLU) system has to handle a large amount of data. A graphical model serves as an advantageous tool for data analysis encoding the dependencies among variables and learning causal relationships. Over the last two decades, the Bayesian network has become a popular representation for encoding uncertain expert knowledge in expert systems. It is an ideal representation for combining prior knowledge; it avoids overfitting of data. Efficient algorithms have been developed for learning Bayesian networks from data, allowing Bayesian networks to be applied to a wide category of problems. In this paper, we give a comprehensive and state-of the-art introduction to the application of Bayesian networks in different aspects of an NLU system, with emphasis on information retrieval. The extensions and variants of Bayesian networks applied to NLU problems have been described. Examples of application examples are given, in order to illustrate the use of Bayesian networks.

Natural language as the basis for meaning representation and inference

Abstract: Semantic inference is an important component in many natural language understanding applications. Classical approaches to semantic inference rely on logical representations for meaning, which may be viewed as being "external" to the natural language itself. However, practical applications usually adopt shallower lexical or lexical-syntactic representations, which correspond closely to language structure. In many cases, such approaches lack a principled meaning representation and inference framework. We describe a generic semantic inference framework that operates directly on language-based structures, particularly syntactic trees. New trees are inferred by applying entailment rules, which provide a unified representation for varying types of inferences. Rules were generated by manual and automatic methods, covering generic linguistic structures as well as specific lexical-based inferences. Initial empirical evaluation in a Relation Extraction setting supports the validity and potential of our approach. Additionally, such inference is shown to improve the critical step of unsupervised learning of entailment rules, which in turn enhances the scope of the inference system. This paper corresponds to the invited talk of the first author at CI-CLING 2008.

GAIML: A new language for verbal and graphical interaction in chatbots

Abstract: Natural and intuitive interaction between users and complex systems is a crucial research topic in human-computer interaction. A major direction is the definition and implementation of systems with natural language understanding capabilities. The interaction in natural language is often performed by means of systems called chatbots. A chatbot is a conversational agent with a proper knowledge base able to interact with users. Chatbots appearance can be very sophisticated with 3D avatars and speech processing modules. However the interaction between the system and the user is only performed through textual areas for inputs and replies. An interaction able to add to natural language also graphical widgets could be more effective. On the other side, a graphical interaction involving also the natural language can increase the comfort of the user instead of using only graphical widgets. In many applications multi-modal communication must be preferred when the user and the system have a tight and complex interaction. Typical examples are cultural heritages applications (intelligent museum guides, picture browsing) or systems providing the user with integrated information taken from different and heterogenous sources as in the case of the iGoogle™ interface. We propose to mix the two modalities (verbal and graphical) to build systems with a reconfigurable interface, which is able to change with respect to the particular application context. The result of this proposal is the Graphical Artificial Intelligence Markup Language (GAIML) an extension of AIML allowing merging both interaction modalities. In this context a suitable chatbot system called Graphbot is presented to support this language. With this language is possible to define personalized interface patterns that are the most suitable ones in relation to the data types exchanged between the user and the system according to the context of the dialogue.

Natural language understanding based on mental image description language Lmd and its application to language-centered robot manipulation

Abstract: The authors have been working on natural language understanding based on the knowledge representation language L md and its application to robot manipulation by verbal suggestion. The most remarkable feature of L md is its capability of formalizing spatiotemporal events in good correspondence with human/robotic sensations and actions, which can lead to integrated computation of sensory, motory and conceptual information. This paper describes briefly the process from text to robot action via semantic representation in L md and the experimental results of robot manipulation driven by verbal suggestion.

Temporal Action Logic for Question Answering in an Adventure Game

Abstract: Inhabiting the complex and dynamic environments of modern computer games with autonomous agents capable of intelligent timely behaviour is a significant research challenge. We illustrate this using our own attempts to build a practical agent architecture on a logicist foundation. In the ANDI-Land adventure game concept players solve puzzles by eliciting information from computer characters through natural language question answering. While numerous challenges immediately presented themselves, they took on a form of concrete and accessible problems to solve, and we present some of our initial solutions. We conclude that games, due to their demand for human-like computer characters with robust and independent operation in large simulated worlds, might serve as excellent test beds for research towards artificial general intelligence.

A Korean Mobile Conversational Agent System

Abstract: This paper presents a Korean conversational agent system in a mobile environment using natural language processing techniques. The aim of a conversational agent in mobile environment is to provide natural language interface and enable more natural interaction between a human and an agent. Constructing such an agent, it is required to develop various natural language understanding components and effective utterance generation methods. To understand spoken style utterance, we perform morphosyntactic analysis, shallow semantic analysis including modality classification and predicate argument structure analysis, and to generate a system utterance, we perform example based search which considers lexical similarity, syntactic similarity and semantic similarity.

Acquiring Commonsense Knowledge about Properties of Concepts from Text

Abstract: Commonsense knowledge plays an important role in various areas such as natural language understanding, information retrieval, etc. This paper presents a method for acquiring commonsense knowledge about properties of concepts by analyzing how adjectives are used with nouns in everyday language. We firstly mine a large scale corpus for potential concept-property pairs using lexico-syntactic patterns and then filter erroneously acquired ones based on heuristic rules and statistical approaches. For each concept, we automatically select the commonsensical properties and evaluate their applicability. Finally, we generate commonsense knowledge represented with explicit fuzzy quantifiers. Experimental results demonstrate the effectiveness of our approach.

Natural Language Understanding

Abstract: We describe several algorithms for developing natural language understanding (NLU) applications. The algorithms include a rule-based system and several statistical systems. We consider two major types of NLU applications: dialog systems and speech mining. For dialog systems, the NLU function aims to understand the full meaning of a userʼs request in the context of a human-machine interaction in a narrow domain such as travel reservation. For speech mining applications, the NLU function aims at detecting the presence of a limited set of concepts and some of their relations in unrestricted human-human conversations such as in a call center or an oral history digital library. We describe in more detail a statistical parsing algorithm using decision trees for dialog systems and two word-tagging algorithms for speech mining.

Research on Intelligent Search Engine Based on Semantic Comprehension

Abstract: This article proposes a search engine model which is based on the natural language understanding It includes a method to analyze users' quest ions in natural language from three layers, that is, keyword, quest ion type and question focus The analysis consists of semantic analysis, feature extraction and semantic matching And with this thought the feature base that faces to Web page content is built In addition, this article proposes an algorithm of returning to the documents arrangement, it investigates implementing retrieval system based on the Lucene toolkit The feature words, which are collected in the feature base, are tested, and the precision ratio is about 867% The test result indicates that the module can realize the semantic comprehension to query, and it has an evident effect to improve the precision of search engine

A Cognitive Substrate for Natural Language Understanding

Abstract: Our goal is to understand human language use and create systems that can use human language fluently. We argue that to a achieve this goal, we must formulate all of the problems for language use from morphology up to pragmatics using the same cognitive substrate of reasoning and representation abilities. We propose such a substrate and described systems based on it. Our arguments, results with real-world systems and ongoing work suggest that the cognitive substrate enables a significant advance in the power of cognitive models and intelligent systems to use human language.

A survey of Chinese text similarity computation

Abstract: There is not a natural delimiter between words in Chinese texts Moreover, Chinese is a semotactic language with complicated structures focusing on semantics Its differences from Western languages bring more difficulties in Chinese word segmentation and more challenges in Chinese natural language understanding How to compute the Chinese text similarity with high precision, recall and low cost is a very important but challenging task Many researchers have studied it for long time In this paper, we examine existing Chinese text similarity measures, including measures based on statistics and semantics Our work provides insights into the advantages and disadvantages of each method, including tradeoffs between effectiveness and efficiency New directions of the future work are discussed

Chatting System, Method And Apparatus For Virtual Pet

Abstract: Embodiments of the present invention provides a chatting system for a virtual pet, including a pet client, configured to receive a natural language sentence of a pet master, and send the natural language sentence to a virtual pet server; the virtual pet server, configured to forward the natural language sentence to a questioning and answering server, and return a natural language response to the pet client; the questioning and answering server, configured to perform processing of natural language understanding on the natural language sentence, obtain language characteristics of the pet master, generate the natural language response according to a natural language understanding result and the language characteristics of the pet master, and return the natural language response to the virtual pet server. Embodiments of the present invention further provide a chatting method for a virtual pet and a questioning and answering server.

Artificial Intelligence: Structures and Strategies for Complex Problem Solving

Abstract: Artificial Intelligence: Structures and Strategies for Complex Problem Solving by George F. Luger 6th edition, Addison Wesley, 2008 The book serves as a good introductory textbook for artificial intelligence, particularly for undergraduate level. It covers major AI topics and makes good connection between different areas of artificial intelligence. Along with each technique and algorithm introduced in the book, is a discussion of its complexity and application domain. There is an attached website to the book that provides auxiliary materials for some chapters, sample problems with solutions, and ideas for student projects. Besides Prolog and Lisp, java and C++ are also used to implement many of the algorithms in the book. The book is organized in five parts. The first part (chapter 1) gives an overview of AI, its history and its various application areas. The second part (chapters 2–6) concerns with knowledge representation and search algorithms. Chapter 2 introduces predicate calculus as a mathematical tool for representing AI problems. The state space search as well as un-informed and heuristic search methods is introduced in chapters 3 and 4. Chapter 5 discusses the issue of uncertainty in problem solving and covers the foundation of stochastic methodology and its application. In chapter 6 the implementation of search algorithms is shown in production system and blackboard architectures. Part 3 (chapters 7–9) discusses knowledge representation and different methods of problem solving, including strong, weak and distributed problem solving. Chapter 7 begins with reviewing the history of evolution of AI representation schemes, including semantic networks, frames, scripts and conceptual graphs. This chapter ends with a brief introduction of Agent problem solving. Chapter 8 presents the production model and rule-based expert systems as well as case-based and model-based reasoning. The methods of dealing with various aspects of uncertainty are discussed in chapter 9. These methods include Dempster-Shafer theory of evidence, Bayesian and Belief networks, fuzzy logics and Markov models. Part 4 is devoted to machine learning. Chapter 10 describes algorithms for symbol-based learning, including induction, concept learning, vision-space search and ID3. The neural network methods for learning, such as back propagation, competitive, Associative memories and Hebbian Coincidence learning were presented in chapter 11. Genetic algorithms and evolutionary learning approaches are introduced in chapter 12. Chapter 13 introduces stochastic and dynamic models of learning along with Hidden Markov Models, Dynamic Baysian networks and Markov Decision Processes. Part 5 (chapters 14 and 15) examines two main application of AI: automated reasoning and natural language understanding. Chapter 14 begins with an introduction to weak methods in problem solving and continues with presenting resolution theorem proving. Chapter 15 deals with the complex issue of natural language understanding by discussing main methods of syntax and semantic analysis of natural language corpus. The chapter ends with examples of natural language application in Database query generation, text summarization and question answering systems. Finally, chapter 16 is a summary of the materials covered in the book as well current AI's limitations and future directions. One criticism about the book would be that the materials are not covered in enough depth. Because of the space limitation, many important AI algorithms and techniques are discussed briefly without providing enough details. As a result, some chapters (e.g., 8, 9, 11, and 13) of the book should be supported by complementary materials to make it understandable for undergraduate students and motivating for graduate students. Another issue is with the structure of the book. The order of presenting chapters introduces sequentially different challenges and techniques in problem solving. Consequently, some topics such as uncertainty and logic are not introduced separately and are distributed in different chapters of the book related to different parts. Although interesting, this makes the book hard to follow. In summary, the book gives a great insight to the readers that want to familiar themselves with artificial intelligence. It covers a broad range of topics in AI problem solving and its practical application and is a good reference for an undergraduate level introductory AI class. Elham S. Khorasani, Department of Computer Science Southern Illinois University Carbondale, IL 62901, USA

Layer structures and conceptual hierarchies in semantic representations for NLP

Abstract: Knowledge representation systems aiming at full natural language understanding need to cover a wide range of semantic phenomena including lexical ambiguities, coreference, modalities, counterfactuals, and generic sentences. In order to achieve this goal, we argue for a multidimensional view on the representation of natural language semantics. The proposed approach, which has been successfully applied to various NLP tasks including text retrieval and question answering, tries to keep the balance between expressiveness and manageability by introducing separate semantic layers for capturing dimensions such as facticity, degree of generalization, and determination of reference. Layer specifications are also used to express the distinction between categorical and situational knowledge and the encapsulation of knowledge needed e.g. for a proper modeling of propositional attitudes. The paper describes the role of these classificational means for natural language understanding, knowledge representation, and reasoning, and exemplifies their use in NLP applications.