This paper summarizes work over the past five years on the automated planning and generation of multisentence texts using discourse structure relations, placing it in context of ongoing efforts by Computational Linguists and Linguists to understand the structure of discourse. Based on a series of studies by the author and others, the paper describes how the orientation of generation toward communicative intentions illuminates the central structural role played by intersegment discourse relations. It outlines several facets of discourse structure relations as they are required by and used in text planners --- their nature, number, and extension to associated tasks such as sentence planning and text formatting. In Artificial Intelligence 63, Special Issue on Natural Language Processing, 1993. This work was partially supported by the Rome Air Development Center under RADC contract FQ7619-8903326 -0001. 1 1 Introduction Every day, people produce thousands of words of connected...

Abstract not found

This report presents a detailed analysis and review of NLP evaluation, in principle and in practice. Part 1 examines evaluation concepts and establishes a framework for NLP system evaluation. This makes use of experience in the related area of information retrieval and the analysis also refers to evaluation in speech processing. Part 2 surveys significant evaluation work done so far, for instance in machine translation, and discusses the particular problems of generic system evaluation. The conclusion is that evaluation strategies and techniques for NLP need much more development, in particular to take proper account of the influence of system tasks and settings. Part 3 develops a general approach to NLP evaluation, aimed at methodologically-sound strategies for test and evaluation motivated by comprehensive performance factor identification. The analysis throughout the report is supported by extensive illustrative examples. This work was carried out under the UK Science and Engineeri...

So That Nobody Has To Go To School If They Don't Want To by Roger Sipher A decline in standardized test scores is but the most recent indicator that American education is in trouble. One reason for the crisis is that present mandatory-attendance laws force many to attend school who have no wish to be there. Such children have little desire to learn and are so antagonistic to school that neither they nor more highly motivated students receive the quality education that is the birthright of every American. The solution to this problem is simple: Abolish compulsory-attendance laws and allow only those who are committed to getting an education to attend. This will not end public education. Contrary to conventional belief, legislators enacted compulsory-attendance laws to legalize what already existed. William Landes and Lewis Solomon, economists, found little evidence that mandatory-attendance laws increased the number of children in school. They found, too, that school systems have never effectively enforced such laws, usually because of the expense involved. There is no contradiction between the assertion that compulsory attendance has had little effect on the number of children attending school and the argument that repeal would be a positive step toward improving education. Most parents want a high school education for their children. Unfortunately, compulsory attendance hampers the ability of public school officials to enforce legitimate educational and disciplinary policies and thereby make the education a good one. Private schools have no such problem. They can fail or dismiss students, knowing such students can attend public school. Without compulsory attendance, public schools would be freer to oust students whose academic or personal behavior undermines the educational mission of the institution. Has not the noble experiment of a formal education for everyone failed? While we pay homage to the homily, "You can lead a horse to water but you can't make him drink," we have pretended it is not true in education. Ask high school teachers if recalcitrant students learn anything of value. Ask teachers if these students do any homework. Quite the contrary, these students know they will be passed from grade to grade until they are old enough to quit or until, as is more likely, they receive a high school diploma. At the point when students could legally quit, most choose to remain since they know they are likely to be allowed to graduate whether they do acceptable work or not. Abolition of archaic attendance laws would produce enormous dividends. First, it would alert everyone that school is a serious place where one goes to learn. Schools are neither day-care centers nor indoor street corners. Young people who resist learning should stay away; indeed, an end to compulsory schooling would require them to stay away. Second, students opposed to learning would not be able to pollute the educational atmosphere for those who want to learn. Teachers could stop policing recalcitrant students and start educating. Third, grades would show what they are supposed to: how well a student is learning. Parents could again read report cards and know if their children were making progress. Fourth, public esteem for schools would increase. People would stop regarding them as way stations for adolescents and start thinking of them as institutions for educating America's youth. Fifth, elementary schools would change because students would find out early they had better learn something or risk flunking out later. Elementary teachers would no longer have to pass their failures on to junior high and high school. Sixth, the cost of enforcing compulsory education would be eliminated. Despite enforcement efforts, nearly 15 percent of the school-age children in our largest cities are almost permanently absent from school. Communities could use these savings to support institutions to deal with young people not in school. If, in the long run, these institutions prove more costly, at least we would not confuse their mission with that of schools. Schools should be for education. At present, they are only tangentially so. They have attempted to serve an all-encompassing social function, trying to be all things to all people. In the process they have failed miserably at what they were originally formed to accomplish.

An approach to connectionist natural language processing is proposed, which is based on hierarchically organized modular Parallel Distributed Processing (PDP) networks and a central lexicon of distributed input/output representations. The modules communicate using these representations, which are global and publicly available in the system. The representations are developed automatically by all networks while they are learning their processing tasks. The resulting representations reflect the regularities in the subtasks, which facilitates robust processing in the face of noise and damage, supports improved generalization, and provides expectations about possible contexts. The lexicon can be extended by cloning new instances of the items, that is, by generating a number of items with known processing properties and distinct identities. This technique combinatorially increases the processing power of the system. The recurrent FGREP module, together with a central lexicon, is used as a ba...

The hierarchical feature map system recognizes an input story as an instance of a particular script by classifying it at three levels: scripts, tracks and role bindings. The recognition taxonomy, i.e. the breakdown of each script into the tracks and roles, is extracted automatically and independently for each script from examples of script instantiations in an unsupervised self-organizing process. The process resembles human learning in that the differentiation of the most frequently encountered scripts become gradually the most detailed. The resulting structure is a hierachical pyramid of feature maps. The hierarchy visualizes the taxonomy and the maps lay out the topology of each level. The number of input lines and the self-organization time are considerably reduced compared to the ordinary single-level feature mapping. The system can recognize incomplete stories and recover the missing events. The taxonomy also serves as memory organization for script-based episodic memory. The map...

A Text-Based Intelligent System should provide more in-depth information about the contents of its corpus than does a standard information retrieval system, while at the same time avoiding the complexity and resource-consuming behavior of detailed text understanders. Instead of focusing on discovering documents that pertain to some topic of interest to the user, an approach is introduced based on the criterion of directionality (e.g., Is the agent in favor of, neutral, or opposed to the event?). A method is described for coercing sentence meanings into a metaphoric model such that the only semantic interpretation needed in order to determine the directionality of a sentence is done with respect to the model. This interpretation method is designed to be an integrated component of a hybrid information access system. 1 Introduction In the light of the increasing availability of computer-accessible full text, an important goal of a Text-Based Intelligent System is to provide a me...

Combinatorial explosion of inferences has always been a central problem in artificial intelligence. Although the inferences that can be drawn from a reasoner's knowledge and from available inputs is very large (potentially infinite), the inferential resources available to any reasoning system are limited. With limited inferential capacity and very many potential inferences, reasoners must somehow control the process of inference. Not all inferences are equally useful to a given reasoning system. Any reasoning system that has goals (or any form of a utility function) and acts based on its beliefs indirectly assigns utility to its beliefs. Given limits on the process of inference, and variation in the utility of inferences, it is clear that a reasoner ought to draw the inferences that will be most valuable to it. This paper presents an approach to this problem that makes the utility of a (potential) belief an explicit part of the inference process. The method is to generate exp...

The paper presents a framework for, and strategies adopted in, an investigation of summarising designed to place future work on automatic summarising on solid foundations. The work reported has been focused on the role of large-scale text structure, and the paper describes comparative studies of different approaches to the characterisation of source text structure and to the use of this structure in summary formation. 2 Introduction In this paper I shall describe some foundationally-motivated work we have been doing in Cambridge, establishing and exploiting a framework for automatic summarising. I shall introduce this account by noting some relations with indexing, and in conclusion return to the connection between the two forms of information capture and use. We can all, and do, summarise; and we use other people's summaries. Summarising, especially in the guise of extracting, was an early task for automation (Luhn, 1958). But not much progress has in fact been made with automatic su...

The availability of on-line corpora is rapidly changing the field of natural language processing (NLP) from one dominated by theoretical models of often very specific linguistic phenomena to one guided by computational models that simultaneously account for a wide variety of phenomena that occur in real-world text. Thus far, among the best-performing and most robust systems for reading and summarizing large amounts of real-world text are knowledge-based natural language systems. These systems rely heavily on domain-specific, handcrafted knowledge to handle the myriad syntactic, semantic, and pragmatic ambiguities that pervade virtually all aspects of sentence analysis. Not surprisingly, however, generating this knowledge for new domain...