The first chapter of this book describes the process of representing the meaning of an input query in the constructed formal language. Following the rule-based semantic analysis for natural langauge questions and definitions, accompanying topics are presented such as the concept of vertical domain, syntactic analysis, first-order language and its metalanguage extensions, as well as the architecture of semantic processor, formal semantics models and grammars relevant to our pragmatic approach.
The second chapter introduces the process of matching the formal representation of an input query with the formalized domain knowledge in the way of semantic headers. The task of query understanding is posed for the hardly-formalized and poorly structured domains with textual answers. Knowledge representation methodology and associated semantic peculiarities are discussed; coding patterns are introduced for the technique of semantic headers.
The third chapter addresses the practical issues of implementing question answering domains. Various topics are discussed from marketing, programming styles and development to deployment and maintenance of question answering applications, as well as management of the team of knowledge engineers, experts and testers. Multiagent implementation, various knowledge domains, the software project structure and development tools are also presented.
The fourth chapter introduces the reasoning mechanism as the background of the suggested approach to Q/A, particularly, scenario-based reasoning about mental attitudes. Default logic is used for correction of the semantic representations.
Dr. Boris Galitsky received his Master Degree in Intelligent Systems and PhD in Computational Linguaistics in 1991 and 1993 respectively in the Institute for Information Transmision Problems of Russian Science Academy. Since then he contributed to multiple NLP-related projects in universities and companies across US, particularly, Internet content provider pioneers in Silicon Valley. Having developed his technology in collaboration with Rutgers University and Artificial Intelligence Lab at MIT, Dr Galitsky founded iAskWeb, the company that was providing question answering services to the leaders of financial industry in US. An author of more than 50 publications in Artificial Intelligence and Bioinformatics, Boris Galitsky is currently a lecturer at Birkbeck-University of London UK.
