Roger C. Schank's research while affiliated with Northwestern University and other places

Publications (100)

Article
Good teaching is good story telling. Case‐based teaching exploits the basic capacity for students to learn from stories and the basic desire of teachers to tell stories that are indicative of their experiences. The premise of software that used case‐based teaching would be to place a student in a situation that the student found interesting and whe...
Chapter
Over many years, Roger Schank and his students developed a series of theories and computer models of natural language processing, guided by two main tenets: that natural language processing cannot be studied in isolation from the purposes for which language is used, and that meaning and world knowledge may be crucial at even the earliest points in...
Article
this paper we will discuss some of the problems which have arisen in the development of parsers, and present a new theory of the way parsing works in the normal reading process. We will describe a program which implements this theory and understands newspaper stories about terrorism
Article
This paper is a first attempt to redress that imbalance; it presents a content-driven approach to indexing and reports preliminary results from our working group at Northwestern University's Institute for the Learning Sciences (ILS). This group's goal has been to devise a content theory for indices of sufficient generality to support the full range...
Article
In the article, the author says that universities are scrambling to get into the distance-education business. There are a few obvious reasons why a university might want to offer virtual courses, reasons ranging from additional student revenue to the exploration of new teaching methods, although the author seriously doubt the latter is high on any...
Article
Full-text available
eader. The crux of the theory is, however, adequately described here. The discussion starts by broadening the current scope of explanation within the field of AI. The much maligned Turing Test is rejected in favour of a test that gauges the level of understanding of a reasoning agent in terms of its ability to generate explanations; not only for ot...
Article
A learning environment architecture called Goal-Based Scenario (GBS) has been designed, which provides motivation, context, and specific challenges, as well as access to information. GBSs are courses structured as a series of scenarios, each involving the pursuit of a clear concrete mission through a set of activities designed to teach a target set...
Article
When considering the use of multimedia systems for education and training there is a hidden danger in the commercial juggernaut of current-generation multimedia software. This danger is that when multimedia are applied to education and training, the information-presentation paradigm that has become so strongly associated with multimedia application...
Article
Margaret Boden's The Creative Mind provides a useful philosopher's-eye perspective on the important subject of creativity. Boden argues that creative thought is necessarily the result of the application of existing knowledge, that the naive notion of creativity is useless, that mechanical creativity is possible, and that the beginning of scientific...
Article
Research in vision and language has traditionally remained separate in part because the classic task of generating a representation of a given image or sentence has resulted in an emphasis on low level structural aspects of these media. In this paper we argue that image and language understanding should be approached with the intent of facilitating...
Article
The Institute for the Learning Sciences is a research and development center established at Northwestern University in 1989 with founding sponsorship from Andersen Consulting. The main goal of the Institute is to create innovative, computer-based learning environments, based on leading-edge research in Cognitive Science and Artificial Intelligence,...
Article
Full-text available
This paper is a first attempt to redress that imbalance; it presents a content-driven approach to indexing and reports preliminary results from our working group at Northwestern University's Institute for the Learning Sciences (ILS). This group's goal has been to devise a content theory for indices of sufficient generality to support the full range...
Article
In this essay, we argue that stories about one's experiences, and the experiences of others, are the fundamental constituents of human memory, knowledge, and social communication. This argument includes three propositions: 1) Virtually all human knowledge is based on stories constructed around past experiences; 2) New experiences are interpreted in...
Article
Outside school, people typically learn during their experiences while addressing desired goals. The Goal-Based Scenario (GBS) framework describes computer-based learning environments that exploit this simple fact. In this article, we propose a structure and a set of design criteria for learn-by-doing environments that enable students to work toward...
Article
As a child, did you learn how to walk by taking a walking class? No. Children learn by doing. Then why force them into passivity in school? Properly designed educational software on multimedia computers supports active participation and puts the student in control. Such software demands the right teaching architecture, as explained here.< >
Book
eader. The crux of the theory is,however, adequately described here.The discussion starts by broadening the current scope of explanation within the fieldof AI. The much maligned Turing Test is rejected in favour of a test that gauges the levelof understanding of a reasoning agent in terms of its ability to generate explanations; notonly for others...
Article
The goals of this project have been to carry out research aimed at implementing and applying case-based reasoning, or CBR (Riesbeck and Schank. 1989) in a variety of distinct tasks and domains. In Particular, the work carried out under this project has focussed on three problems: (1) The development of a robust memory-based parsing technology (Dire...
Article
Four viewpoints about what artificial intelligence is about are surveyed. Schank describes a program exhibiting AI as one that can change as a result of interactions with the user. Such a program would have to process hundreds or thousands of examples as opposed to a handful. Because AI is a machine's attempt to explain the behavior of the (human)...
Article
We view the future of artificial intelligence from three perspectives: scientific, technological, and educational. The common theme among these views is the central importance of learning, particularly in problem domains where case-based reasoning seems appropriate. From a scientific perspective, an adequate model of the mind must account for the p...
Article
Most efforts in making computer-based teachers have not tried to change the standard teaching methods commonly used in schools today to any significant degree. We survey several of these methods of teaching and examine both their strengths and weaknesses. We then introduce six new teaching architectures, defined as a kind of general blueprint or fr...
Conference Paper
Creativity is obviously a crucial aspect of human intelligence, and yet it has not been explored much by AI researchers. A principal reason for this lack of attention is the mystical aura that the word “creativity” has about it. Creativity is thought to be something so mysterious that AI researchers have been intimidated by it. We claim that creati...
Article
Explanation-based learning (EBL) is a very powerful method for category formation. Since EBL algorithms depend on having good explanations, it is crucial to have effective ways to build explanations, especially in complex real-world situations where complete causal information is not available. When people encounter new situations, they often expla...
Article
We have the following goals for our research with the direct memory access algorithm for understanding and inference: (a) using the DMA algorithm to carry out larger scale case-based reasoning; (b) improving the robustness of the understander; (c) exploring issues in parallelizing the algorithm. In what follows below, we will not make detailed refe...
Conference Paper
Connectionist models inherently include features and exhibit behaviors which are difficult to achieve with traditional logic-based models. Among the more important of such characteristics are 1) the ability to compute nearest match rather than requiring ...
Article
Traducción de: Scripts Plans, Goals and Understanding Obra en la intersección de la psicología evolutiva con la inteligencia artificial, cuya intención es abordar la naturaleza del conocimiento y la forma como es usado.
Article
The inductive category formation framework, an influential set of theories of learning in psychology and artificial intelligence, is deeply flawed. In this framework a set of necessary and sufficient features is taken to define a category. Such definitions are not functionally justified, are not used by people, and are not inducible by a learning s...
Article
Full-text available
The concept of artificial intelligence (AI) is explored and conceptial structures are sought that are needed to make computers perform operations natural to the human mind. AI is considered to be a much more basic science than is popularly believed. Many best AI ideas require a great deal of work before they can become useful applications. When an...
Conference Paper
This paper describes a natural language generator program that has knowledge about social and psychological aspects of conversation such as relative social statuses of speaker and hearer, models of the various hearers, some perception of the hearers' interests and sympathies, as well as its own purposes with the conversation. The program contains a...
Article
In this paper we discuss one approach of artificial intelligence (AI) research on natural language processing that we believe will shed light on questions about coherence. Specifically, we think it will help us understand the factors that determine whether an understander will find a given text or input coherent or incoherent. The central theme of...
Chapter
My aim here is to attempt to determine some of the rules that may legitimately be used for the process of response creation in a conversational setting. My assumption is that people bring a great deal of background information and rules about how to interpret what someone really means to the conversational setting. The process of communication is e...
Article
We present a theory of conversation comprehension in which a line of the conversation is “understood” by relating it to one of seven possible “points”. We define these points, and present examples where it seems plausible that the failure to “get the point” would indeed constitute a failure to understand the conversation. We argue that the recognit...
Article
An abstract is not available.
Article
This paper explores the role of syntax in computational theories of natural language. We discuss the integrated processing hypothesis, which contends that meaning and world knowledge play a crucial part in language understanding even at the earliest points in the process. The hypothesis implies that syntactic knowledge plays no privileged role in l...
Article
This paper outlines some of the issues and basic philosophy that have guided my work and that of my students in the last ten years. It describes the progression of conceptual representational theories developed during that time, as well as some of the research models built to implement those theories. The paper concludes with a discussion of my mos...
Article
This paper outlines some of the issues and basic philosophy that have guided my work and that of my students in the last ten years. It describes the progression of conceptual representational theories developed during that time, as well as some of the research models built to implement those theories. The paper concludes with a discussion of my mos...
Article
Various levels of understanding in story comprehension are presented. Five levels, Event Characterization, Event Connection, Contextual Understanding, International Understanding, and Global-Contextual Understanding, applicable to both computer and human understanding are introduced. Question answering and paraphrase are examined as two methods by...
Chapter
This chapter presents the seven basic social acts and the triangle representation as a very useful set of tools in analyzing sociopolitical interactions. Each social act groups a set of inferences about the social situation it represents. A very large number of social conflicts, actions, and situations can be represented by combinations of this sma...
Article
: The problem of controlling inference is one of the most serious in Artificial Intelligence. New types of goal and plan inferences seriously compound the problem. This paper attempts to outline one possible solution to controlling inferences, namely following what is interesting and ignoring what is not. (Author)
Conference Paper
This paper outlines a strategy for conversational analysis that utilizes multiple levels of information flow in order to characterize the conceptual content of a conversation. The role of predictive knowledge structures and contextual information is illustrated in an analysis of a sample conversation, along with a partial outline of conversational...
Conference Paper
A new type of natural language parser is presented. The idea behind this parser is to map input sentences into the deepest form of the representation of their meaning and make appropriate inferences during the parsing process, using interest to guide the processing.
Article
The problem of how to organize particular experiences in a long-term memory has been largely neglected in Natural Language Understanding research. This paper addresses that problem. It shows why good memory organization is necessary for doing intelligent tasks such as story understanding and conversation. A computer memory organization modeled afte...
Article
A new type of natural language parser is presented. The idea behind this parser is to map input sentences into the deepest form of the representation of their meaning and inferences, as is appropriate. The parser is not distinct from an entire understanding system. It uses an integrated conception of inferences, scripts, plans, and other knowledge...
Article
For both people and machines, each in their own way, there is a serious problem in common of making sense out of what they hear, see, or are told about the world. The conceptual apparatus necessary to perform even a partial feat of understanding is formidable and fascinating. Our analysis of this apparatus is what this book is about. —Roger C. Scha...
Article
The state of research in automatic natural language processing at Yale University is described. Five programs are discussed: SAM, FRUMP, PAM, TALESPIN, and POLITICS.
Article
A natural language utterance may denote an attitude--a disposition of a person toward another person, object, activity, or state. In order to represent the meaning of such utterances, the writers propose a small set of primitive attitudinal states. Sentences which denote attitudes can then be represented in terms of these attitudinal primitives. Th...
Article
An abstract is not available.
Chapter
A frequently ignored issue in the disciplines of psychology and linguistics is the structure of texts. Sentences rarely come in isolation, but rather in groups, the meaning of whose combination is more than the meaning of the individual parts. Such meaning combination can occur within a sentence as well.
Article
Rules of conversation are given that specify what can follow what. A system for deciding what makes a reasonable subject for a conversation is shown. Topics are discussed and rules for topic shift are presented.
Article
Using knowledge to make inferences is a crucial aspect of natural language processing. While the notion of a script has been useful in this regard, other forms of knowledge are also needed. In particular, knowledge about goals and knowledge about stories is crucial for story understanding. The understander must know what types of goals characters i...
Conference Paper
,This paper,discusses ,the ,kind ,of information,that ,must ,be present ,in a computer program,that,models ,the ,linguistic ,development,of achild.,A three stage model ,is presented ,that characterizes,the ,development,of a ,natural ,lan­ guage parser in a child of ages one, one and a half, and two. Some data from children of these ages,is presente...
Conference Paper
This paper is a condensed version of the author's thesis [Bolles 1976], which investigates a subclass of visual information processing referred to as verification vision (abbreviated VV). VV uses a model of a scene to locate objects of interest in a ...
Article
The authors respond to Dresher and Hornstein's article (EJ 161 384, Cognition, December, 1976) on artificial intelligence (AI). The dispute between linguistic theorists and AI researchers is based upon their different aims; while AI researchers develop programs capable of intelligent behavior, transformational linguists study the characteristics of...
Article
I have so far managed to keep out of the great Weizenbaum debate, although I find my name and research bandied about enough. However, the last piece by Weizenbaum written in response to McCarthy's SIGART review has caused me to take the gauntlet at last.
Article
The MARGIE system is a set of three programs that attempt to understand natural language. They are based on the Conceptual Dependency system for meaning representation. The analysis program maps sentences into conceptual structures. The memory program makes inferences from input conceptual structures. The generator codes conceptual structures back...
Article
An abstract is not available.
Article
The treatment of a certain class of adverbs in conceptual representation is given. Certain adverbs are shown to be representative of complex belief structures. These adverbs serve as pointers that explain where the sentence that they modify belongs in a belief structure.
Article
We examine the original goals of generative linguistic theory. We suggest that these goals were well defined but misguided with respect to their avoidance of the problem of modelling performance. With developments such as Generative Semantics, it is no longer clear that the goals are clearly defined. We argue that it is vital for linguistics to con...
Article
The notion of computer understanding of natural language is examined relative to inference mechanisms designed to function in a language-free deep conceptual base (Conceptual Dependency). The conceptual analysis of a natural language sentence into this conceptual base, and the nature of the memory which stores and operates upon these conceptual str...
Conference Paper
A program is described that accepts natural language input and makes inferences from it and paraphrases of it. The Conceptual Dependency framework is the basis of this system.
Article
Reviews research undertaken to create a theory of human natural language understanding. It is noted that those who have made attempts to solve the problem have had to restrict the domain of the particular problem that they were trying to solve, and sacrifice theoretical considerations for programming considerations. Psychiatric interviewing program...
Article
In order to create conceptual structures that will uniquely and unambiguously represent the meaning of an utterance, it is necessary to establish 'primitive' underlying actions and states into which verbs can be mapped. The paper presents analysis of the most common mental verbs in terms of such primitive actions and states. In order to represent t...
Conference Paper
The conceptual dependency analyzer described in the first IJCAI (8) has been modified so as to function more conceptually with less reliance on syntactic rules. In order to have an analyzer be conceptually driven, it is necessary for the sys­ tem to know what it is looking for. That is, it must make predictions as to what can follow con­ ceptually...
Conference Paper
This paper describes an operable automatic parser for natural language. It is a conceptual parser, concerned with determining the underlying meaning of the input utilizing a network of concepts explicating the beliefs inherent in a piece of discourse.
Article
This paper describes an operable automatic parser for natural language. The parser is not concerned with producing the syntactic structure of an input sentence. Instead, it is a conceptual parser, concerned with determining the underlying meaning of the input. Given a natural language input, the parser identifies and disambiguates the concepts deri...
Article
Schank argues that educational systems (for schools or business)should be redesigned so they consist of goal-based scenarios (GBS). The intent of a GBS is to provide motivation, a sense of accomplishment and a support system for the student, along with a focus on skills rather than facts. Goal-based scenarios allow students to pursue well-defined g...
Article
~SrRACT. The MARGIE system is a set of three programs that attempt to understand natural language. They are based on the Conceptual Dependency system for meaning representation. The analysis program maps sentences into conceptual structures. The memory program makes inferences from input conceptual structures. The generator codes conceptual structu...

Citations

... The Case-Based Reasoning (CBR) method was originally pioneered by Schank [30] from his research work on dynamic memory [31], in which he discovered that memory recalls solved cases as similarity patterns to solve new problems posed. In this way, he defines reasoning based on cases as "the resolution of problems using or adapting the solutions of old problems" [32]. Case-Based Reasoning (CBR) is an artificial intelligence paradigm that combines problem solving and learning, as established by Watson [33]-that is, Case-Based Reasoning (CBR) is a method to describe a problem posed where various technologies can be used for its resolution. ...
... It is not surprising that research on comprehension started at the other end-where global structures could be seen as shaping local processes. Thus, early AI systems designed for comprehension started with global organizers, conceptual scripts for restricted situational comprehension (Schank & Abelson, 1977). Similarly, approaches within psychology and education emphasized schemata, situated conceptual structures (Anderson & Pearson, 1984). ...
... The representations of such complex actions have also been referred to as 'scripts': for example the script of how to make breakfast which in turn consists of boiling tea and preparing bread, each of which can be further broken down into subcomponent actions. 3 Such hierarchical and temporally extended action representations are being stored in memory and employed in understanding or preparing for new experiences (Schank, 1980;Schank and Abelson, 1975). Moreover, the action representations that are contained in scripts can be recomposed in manifold ways and employed in different kinds of cognitive processes, ranging from the immediate performance of an action to imagination of potential future actions or understanding the actions by other agents (Cooper and Shallice, 2006;Norman and Shallice, 1986). ...
... É curioso observar a atitude semelhante em relação ao conhecimento lingüístico existente entre os estudiosos da inteligência artificial. R. SCHANK (1982;, por exemplo, afirma que "as regras gramaticais não são decisivas para a compreensão" e que as decisivas mesmo são as "regras" da realidade". ...
... As Schunk (1978) stated, the above theories are common in some part such as learners have been involved in learning and structuring solutions so that they can get the maximum amount of information. Constructivists such as Piaget and Vygostky suggest that through active participation and by engaging in social interaction more deeply, meaningful learning can be achieved (Palinscar, 1998). ...
Reference: Full Issue
... Production schemata have been defined for various T genres--e.g. narratives; argumentations in requests and attitude expression; expository texts to answer factual questions; news stories produced by the computer and by humans ( van Dijk, 1985;Hovy & Schank, 1984;McKeown, 1985;Zammuner, 1987a;Zammuner and Boscolo, 1988). When prototypical schemata exist, they can function as a procedural aid because P can approach linearisation issues and text planning by relying on a known schema. ...
... Par la suite, des approches sémantiques sont peu à peu proposées et intégrées dans des systèmes de plus en plus modulaires mais qui continuent encore à ne s'appliquer qu'en domaine restreint. On voit désormais des modèles linguistiques intégrer les systèmes comme la théorie de Schank des dépendances conceptuelles [132], implémentée pour la première fois en 1973 dans MARGIE [135], puis réutilisée dans SAM [133] et PAM [155]. Les systèmes de l'époque utilisent des structures de connaissances dépendantes du domaine d'application et écrites à la main par des experts du domaine sous forme de règles et de bases de données. ...
... • Autres algorithmes: Forward-Backward sampling (Fung, 1994), Self importance (Shachter, 1989), Heuristic importance (Shachter, 1989 (Schank, 1982) (DARPA, 1991) (Schank, 1990) (Kolodner, 1993). CBR est une combinaison de connaissances et de processus de gestion et de réutilisation d'expériences passées. ...
... Aristotle commented on this characteristic of riddle in the Poetics (Kenny 2013) that "the very nature indeed of a riddle is this, to describe a fact in an impossible combination of words". Commonsense reasoning and creative thinking (Schank and Kass 1987) are required in order to develop a valid explanation for the seemingly impossible description of riddling object. We believe that riddles, as a thought provoking challenge for human, can help sharpen deeper language understanding and reasoning abilities of machines as well. ...
... 5. Knowledge machine should have an ability to take into account not only descriptions of situations in its memory but results as well, providing a possibility to reconstruct a description from a result, i.e. if you told it a result, it would be able to evolve what the steps were which led up to that result. 6. Possessing information about the great number of cases, knowledge machine should have an ability to find a strong family resemblance about them, i.e. to find templates of typical cases. ...