Showing papers on "Question answering published in 1987"

Disambiguating prepositional phrase attachments by using on-line dictionary definitions

Abstract: Standard on-line dictionaries offer a wealth of knowledge expressed in natural language form. We claim that such knowledge can and should be accessed by natural language processing systems to solve difficult ambiguity problems. This paper sustains that claim by describing a set of computational tools and techniques used to disambiguate prepositional phrase attachments in English sentences, by accessing on-line dictionary definitions. Such techniques offer hope for eliminating the time-consuming, and often incomplete, hand coding of semantic information that has been conventional in natural language understanding systems.

Knowledge systems and PROLOG

Abstract: Every word to utter from the writer involves the element of this life. The writer really shows how the simple words can maximize how the impression of this book is uttered directly for the readers. Even you have known about the content of knowledge systems and prolog so much, you can easily do it for your better connection. In delivering the presence of the book concept, you can find out the boo site here.

A prototype of an intelligent system for information retrieval: IOTA

Abstract: Recent results in artificial intelligence research are of prime interest in various fields of computer science; in particular we think information retrieval may benefit from significant advances in this approach. Expert systems seem to be valuable tools for components of information retrieval systems related to semantic inference. The query component is the one we consider in this paper. IOTA is the name of the resulting prototype presented here, which is our first step toward what we call an intelligent system for information retrieval . After explaining what we mean by this concept and presenting current studies in the field, the presentation of IOTA begins with the architecture problem, that is, how to put together a declarative component, such as an expert system, and a procedural component, such as an information retrieval system. Then we detail our proposed solution, which is based on a procedural expert system acting as the general scheduler of the entire query processing. The main steps of natural language query processing are then described according to the order in which they are processed, from the initial parsing of the query to the evaluation of the answer. The distinction between expert tasks and nonexpert tasks is emphasized. The paper ends with experimental results obtained from a technical corpus, and a conclusion about current and future developments.

Multi-window displays for readers of lengthy texts

Abstract: Two experiments explore whether it would help readers re-locate information in an “electronic book” if different windows on the screen were used to display specific sections of the text. Experiment 1, using a within-subject design, showed that reading and question answering were faster with a single window than with a multi-window display. Experiment 2, in which procedural skills were developed before starting the experiment, and a between-subject design was used, showed that this advantage for the single window display would not generally be the case. The multi-window display was a significant help to readers relocating information once they were familiar with the procedures for manipulating the text. The studies suggest ways in which the display of lengthy electronic texts may be improved. They also illustrate the ease with which misleading results can be obtained in studies of human-computer interaction, and emphasize the need for establishing adequate levels of procedural skill before exploring display characteristics.

Conceptual retrieval and case law

Abstract: The main body of the paper gives a rundown of the research undertaken for my doctoral dissertation the design of a mode1 for the retrieval of of law cases, with emphasis on the development of a knowledge representation. The project intersects a number of distinct interest areas: information retrieval, text processing, artificial intelligence, and legal reasoning. In section 2, the areas of intersection are defined.

Combining discourse strategies to generate descriptions to users along a naive/expert spectrum

Abstract: A question answering system that provides access to a large amount of data will be most useful if it can tailor its answer to each user. In particular, a user's level of knowledge about the domain of discourse is an important factor in this tailoring. In previous work we determined that a user's level of domain knowledge affects the kind of information provided in an answer to a user's question as opposed to just the amount of information, as was previously proposed. We also explained how two distinct discourse strategies could be used to generate texts aimed at naive and expert users. Users are not necessarily truly expert or fully naive however, but can be anywhere along a knowledge spectrum whose extremes are naive and expert In this work, we show how our generation system, TAILOR, can use information about a user's level of expertise to combine several discourse strategies in a single text, choosing the most appropriate at each point in the generation process, in order to generate texts for users anywhere along the knowledge spectrum. TAILOR'S ability to combine discourse strategies based on a user model allows for the generation of a wider variety of texts and the most appropriate one for the user.

Applied natural language processing

Abstract: As far as successful engineering of artificial intelligence concepts and their commercial impact is concerned, expert system technology has attracted the interest of those people currently involved in the AI business. Shwartz's book, however, deals with an entirely different stream of AI applications which he thinks is ready for the market - natural language processing. The volume is based on the author's experience on development teams at Cognitive Systems and Intelligent Business Systems. These companies were created to take basic research ideas from Yale University and make them into robust, well-engineered natural language systems. And that's what the book is basically about. It is intended neither to give a comprehensive overview of the various application sides of natural language research, nor to consider alternative research methodologies, but frankly limits its scope to the Schankian paradigm of natural language understanding and the feasibility of its transition into the world of commercial natural language technology. As a consequence, the book is deliberately "oriented toward a business, rather than an academic audience, [namely] corporate strategic planners, data processing and information center personnel" [p.xxiii], and all those involved in the "computer industry who need to evaluate and understand the applicability of this technology within their organizations" [idem.]. The book is organized into ten chapters. The first two cover the theoretical underpinnings of the emerging natural language technology. Chapter 1 provides the methodological foundations for the distinction between natural language processing and natural language understanding. It considers the role of world knowledge (knowledge representation structures and associated reasoning facilities) to be the key element of human language understanding. Chapter 2 picks up this point by providing a basic and lucidly written introduction into Schank's Conceptual Dependency theory. Emphasis is on the version elaborated in Schank (1975) and its application in En-type conceptual parsers. Follow-up constructs developed in later phases within that paradigm are treated in considerably less detail (e.g., scripts, plans and goals), or are even not considered at all (e.g. memory organization packets (MOPs), etc.). However, major principles of conceptual natural language analysis are comprehensively laid down: the status of semantic primitives as abstract, language-independent representational entities, the intimate coupling of knowledge structures and the associated reasoning

Natural Language Processing: A Knowledge-Engineering Approach

Abstract: This book is long overdue. With the recent explosion of interest in Al, technology has been advancing at a dizzying rate. Much like a runaway bull market, it's easy to get caught up in momentary enthusiasm and think that progress has no limit. And like investors transported in a stock market craze, we have ignored many important fundamentals, thinking that somehow this time is different. But is it different? Can nothing but good for all come from progress in Al? Is technology really value-neutral and culturally unbiased, or is it a subtle form of western imperialism? Why has Al refused to stand up to the same rigorous standards as other sciences? Are we really creating intelligence-human intelligence-or just a small subset with a clever facade? The 27 essays in this book raise these and many other fascinating questions. Ajit Narayanan opens the section on philosophical issues with the wellthought-out article \"Why Al Cannot Be Wrong.\" Narayanan wants \"to demonstrate that unless Al is provided with a proper theoretical basis and an appropriate methodology, one can say just about anything one wants to about intelligence and not be contradicted.\" He further asserts that \"unless Al is provided with some reasonable goals and objectives, little of current Al research can be said to be progressing.\" Taking a tough stand, he argues his point well. In \"The Effect of IT [information technology] on Women's Lives,\" Ursula Huws argues that information technology may-if it does not alreadyadversely and disproportionately affect women and minorities. \"Many jobs which have been traditionally defined as highly skilled and carried out exclusively by white men have become 'deskilled' . . they have been simplified, casualized and opened up for women and for people from ethnic minority groups . . . \" Such new opportunities for employment may not be great opportunities after all. Steve Torrance points out in \"Ethics, Mind and Artifice\" that serious ethical questions are raised if intelligent machines achieve genuine intentional states such as consciousness, rather than token programmed states. \"Genuinely intelligent and sentient artifacts buried beneath the rubble of an earthquake, for instance, would, if still 'alive; no doubt have a direct claim to be rescued, just as would human or animal victims.\" An interesting point-will we require expert systems to pass the Turing test before we agree to save them? In his critique of current educational thinking regarding technology, \"IT, Al and the Electronic Sabre-Tooth, \" David Smith underscores the theme running through many of these articles. He proclaims that \"Any curriculum which does not address itself to fundamental issues will fail-and the price of failure could be social catastrophe.\" And he goes on to comment that many curricula being offered \"have little relevance to the world of today, much less the world of the day after tomorrow.\" Certainly, technology has been advancing for centuries. Why (all of a

An approach to natural language for document retrieval

Abstract: Document retrieval systems have been restricted, by the nature of the task, to techniques that can be used with large numbers of documents and broad domains. The most effective techniques that have been developed are based on the statistics of word occurrences in text. In this paper, we describe an approach to using natural language processing (NLP) techniques for what is essentially a natural language problem - the comparison of a request text with the text of document titles and abstracts. The proposed NLP techniques are used to develop a request model based on “conceptual case frames” and to compare this model with the texts of candidate documents. The request model is also used to provide information to statistical search techniques that identify the candidate documents. As part of a preliminary evaluation of this approach, case frame representations of a set of requests from the CACM collection were constructed. Statistical searches carried out using dependency and relative importance information derived from the request models indicate that performance benefits can be obtained.

A knowledge framework for natural language analysis

Abstract: Recent research in language analysis and language generation has highlighted the role of knowledge representation in both processes. Certain knowledge representation foundations, such as structured inheritance networks and feature-based linguistic representations, have proved useful in a variety of language processing tasks. Augmentations to this common framework, however, are required to handle particular issues, such as the ROLE RELATIONSHIP problem: the task of determining how roles, or slots, of a given frame, are filled based on knowledge about other roles. Three knowledge structures are discussed that address this problem. The semantic interpreter of an analyzer called TRUMP (TRansportable Understanding Mechanism Package) uses these structures to determine the fillers of roles effectively without requiring excessive specialized information about each frame.

Knowledge Systems and Prolog: A Logical Approach to Expert Systems and Natural Language Processing

Abstract: This book is about knowledge systems and Prolog. By 'knowledge systems' the authors mean expert systems, natural language understanding systems, and the combination of the two. According to the authors, the book caters to several groups of readers, viz. those who are interested in knowledge systems, those wishing to familiarize themselves with Prolog programming, and those who are already Prolog programmers but would like to take a more in-depth look at this language. Different chapter sequences are recommended for each group. The introduction to Prolog naturally invites comparisons with more introductory texts, e.g., Clocksin and Mellish, 1981. The book under review covers many of the basic topics, sometimes somewhat more in-depth than Clocksin and Mellish. Especially for those who wish to have a self-contained work on Prolog and its major applications, this comprehensiveness may be an advantage, although many of those curious about knowledge systems and the use of logicbased database query techniques are likely to be familiar with at least the basis of Prolog already. I will leave the knowledge system applications undiscussed here, and concentrate for a moment on the natural language understanding part. The book's main topic as far as natural language processing is concerned is the querying of structured databases in English. Two comments are in order here. In the first place, the book has a flavor of definiteness about results in natural language processing that is not wholly appropriate. As anyone who has worked in the field, particularly those who have worked on large grammars for computers, will know, natural language is notoriously hard to process completely. In the second place, even assuming that every possible query in English is translatable into some logical language according to a wholly formal process, there is still no answer to problems of linguistic performance. That is, language use allows for an almost unlimited variety in paraphrasing statements and questions, and writing a grammar for all these possible paraphrases would most likely lead to the

Answering queries on embedded-complete database schemes

Abstract: It has been observed that, for some database schemes, users may have difficulties retrieving correct information, even for simple queries. The problem occurs when some implicit “piece” of information, defined on some subset of a relation scheme, is not explicitly represented in the database state. In this situation, users may be required to know how the state and the constraints interact before they can retrieve the information correctly.In this paper, the formal notion of embedded-completeness is proposed, and it is shown that schemes with this property avoid the problem described above. A polynomial-time algorithm is given to test whether a database scheme is independent and embedded-complete. Under the assumption of independence, it is shown that embedded-complete schemes allow efficient computation of optimal relational algebra expressions equivalent to the X-total projection, for any set of attributes X.

Some problems with prolog as a knowledge representation language for legal expert systems

Abstract: (1987). Some problems with prolog as a knowledge representation language for legal expert systems. International Review of Law, Computers & Technology: Vol. 3, No. 1, pp. 52-67.

A representation for natural category systems

Abstract: Most AI systems model and represent natural concepts and categories using uniform taxonomies, in which no level in the taxonomy is distinguished. We present a representation of natural taxonomies based on the theory that human category systems are non-uniform. There is a basic level which forms the core of a taxonomy; both higher and lower levels of abstraction are less important and less useful. Empirical evidence for this theory is discussed, as are the linguistic and processing implications of this theory for an artificial intelligence/natural language processing system. Among these implications are: (1) when there is no context effect, basic level names should be used; (2) systems should identify objects as members of their basic level categories more rapidly than as members of their superordinate or subordinate categories. We present our implementation of this theory in SNePS, a semantic network processing system which includes an ATN parser-generator, demonstrating how this design allows our system to model human performance in the natural language generation of the most appropriate category name for an object. The ability of our system to acquire classificational information from natural language sentences is also demonstrated.

A logic for information retrieval

Abstract: This paper examines the potential of recent work in artificial intelligence for the development of more effective information retrieval systems. The primary task in this research has been to examine and define the role of an expert system in the domain of bibliographic retrieval. Once such a goal can be described the available knowledge representations and techniques can be evaluated. This paper examines the role of an expert bibliographic retrieval system, examines an artificial intelligence view of information retrieval, and then describes a prototype expert information retrieval system that has been designed and implemented.

A generative grammar approach for the morphologic and morphosyntactic analysis of Italian

Abstract: A morphologic and morphosyntactic analyzer for the Italian language has been implemented in VM/Prolog |3| at the IBM Rome Scientific Center as part of a project on text understanding.Aim of this project is the development of a prototype which analyzes short narrative texts (press agency news) and gives a formal representation of their "meaning" as a set of first order logic expressions. Question answering features are also provided.The morphologic analyzer processes every word by means of a context free grammar, in order to obtain its morphologic and syntactic characteristics.It also performs a morphosyntactic analysis to recognize fixed and variable sequences of words such as idioms, date expressions, compound tenses of verbs and comparative and superlative forms of adjectives.The lexicon is stored in a relational data base under the control of SQL/DS [2], while the endings of the grammar are stored in the workspace as Prolog facts.A friendly interface written in GDDM [1] allows the user to introduce on line the missing lemmata, in order to directly update the dictionary.

Information retrieval from never-ending stories

Abstract: The System for Conceptual Information Summarization, Organization, and Retrieval (SCISOR) is a research system that consists of a set of programs to parse short newspaper texts in the domain of corporate takeovers and finance. The conceptual information extracted from these stories may then be accessed through a natural language interface. Events in the world of corporate takeovers unfold slowly over time. As a result of this, the input to SCISOR consists of multiple short articles, most of which add a new piece of information to an ongoing story. This motivates a natural language, knowledge-based approach to information retrieval, as traditional methods of document retrieval are inappropriate for retrieving multiple short articles describing events that take place over time. A natural language, knowledge-based approach facilitates obtaining both concise answers to straightforward questions and summaries or updates of the events that take place. The predictable events that take place in the domain make expectation-driven, partial parsing feasible.

Expert systems and information retrieval

Abstract: The existing bibliographic retrieval systems are too complex to permit direct on-line access by untrained end users. Expert system approaches have been introduced in the hope of simplifying the document indexing, search and retrieval operations and rendering these operations accessible to end users. The expert system approach is examined briefly in this note and the conclusion is reached that expert systems are unlikely to provide much relief in ordinary retrieval environments. Simpler and more effective retrieval systems than those currently in use can be implemented by falling back on methodologies proposed and evaluated over twenty years ago that operate without expert system intervention.

Spontaneous retrieval in a conceptual information system

Abstract: A traditional paradigm for retrieval from a conceptual knowledge base is to gather up indices or features used to discriminate among or locate items in memory, and then perform a retrieval operation to obtain matching items These items may then be evaluated for their degree of match against the input This type of approach to retrieval has some problems It requires one to look explicitly for items in memory whenever the possibility exists that there might be something of interest there Also, this approach does not easily tolerate discrepancies or omissions in the input features or indices In a question-answering system, a user may make incorrect assumptions about the contents of the knowledge base This makes a tolerant retrieval method even more necessary An alternative, two-stage model of conceptual information retrieval is proposed The first stage is a spontaneous retrieval that operates by a simple marker-passing scheme It is spontaneous because items are retrieved as a by-product of the input understanding process The second stage is a graph matching process that filters and evaluates items retrieved by the first stage This scheme has been implemented and validated in the SCISOR information retrieval system

Generation: a new frontier of natural language processing?

Abstract: Comprehension and generation are the two complementary aspects of natural language processing (NLP). However, much of the research in NLP until recently has focussed on comprehension. Some of the reasons for this almost exclusive emphasis on comprehension are (1) the belief that comprehension is harder than generation, (2) problems in comprehension could be formulated in the AI paradigm developed for problems in perception, (3) the potential areas of applications seemed to call for comprehension more than generation, e.g., question-answer systems, where the answers can be presented in some fixed format or even in some nonlinguistic fashion (such as tables), etc. Now there is a flurry of activity in generation, and we are definitely going to see a significant part of future NLP research devoted to generation. A key motivation for this interest in generation is the realization that many applications of NLP require that the response produced by a system must be flexible (i.e., not produced by filling in a fixed set of templates) and must often consist of a sequence of sentences (i.e., a text) which must have a textual structure (and not just an arbitrary sequence of sentences containing the necessary information). As the research in generation is taking roots, a number of interesting theoretical issues have become very important, and these are likely to determine the paradigm of research in this "new" area.

A logic-based information modelling language

Abstract: A user-friendly language INFOMOD, intended for information modelling and information systems specification, is described. The language, based on predicate logic, fulfills the role of an implementation-independent Conceptual Schema Language. Both static and dynamic aspects of information systems are specifiable in the language. The language has been in practical use for three years, and has been found suitable for many different kinds of systems, such as conventional database systems, computer-aided manufacturing systems, special purpose information systems, and some expert systems.

The natural language metaphor: an approach to avoid misleading expectations

Abstract: Natural language is often used as a query language for data bases. Up to now only subsets with restricted syntax and vocabulary in definite domains are implemented (often as prototypes). In spite of such restrictions, these query interfaces create a “natural language metaphor”, which supports, but may mislead users. To overcome these problems, we investigate wether we may preserve the natural language metaphor, but prevent the user of misapplying it. Our approach is to specify the global metaphor in a set of more exact analogies and to supplement natural language input by other interaction techniques such as menus or direct manipulation. We do not stress the problems of natural language processing, but concentrate on the relevant metaphors. Our investigation is connected with the development of a natural language query prototype system for knowledge bases in the office domain.

A strategy for reasoning in natural language

Abstract: It is commonly believed that "intelligent" manipulation of natural language (NL) requires the translation of texts into some internal form (e.g., deep structure). However, many disadvantages also arise from the use of internal forms. In order to avoid most of them, reasoning directly on texts should be considered. But this alternative has its own drawbacks and is not generally taken seriously since it entails a dramatic increase in the already “explosive” nature of the process. We discuss here a pattern-matching technique and a strategy, called caricature, that are shown to counteract the effect of this explosion, and we give some results and prospects.