Skip to main content

Paul O'Rorke

Followers

7

Following

1

Co-author

1

Public Views

Interests

Uploads

Papers by Paul O'Rorke

Generalization for explanation-based schema acquisition

National Conference on Artificial Intelligence, Aug 6, 1984

This paper is about explanation-based learning for heuristic problem solvers which "build" soluti... more This paper is about explanation-based learning for heuristic problem solvers which "build" solutions using schemata (frames like scripts) as both "bricks" and "mortar". The heart of the paper is a description of a generalization method which is designed to extract as much information as possible from examples of successful problem solving behavior. A' related generalizer, (less powerful but more efficient), has been implemented as part of an experimental apprentice.*

Focusing on the most important explanations: decision-theoretic horn abduction

National Conference on Artificial Intelligence, Aug 1, 1994

This paper describes a new method, called Decision-Theoretic Horn Abduction (DTHA), for generatin... more This paper describes a new method, called Decision-Theoretic Horn Abduction (DTHA), for generating and focusing on the most important explanations. A procedure is given that can be used iteratively to generate a sequence of explanations from the most to the least important. The new method considers both the likelihood and utility of partial explanations and is applicable to a wide range of tasks. This paper shows how it applies to an important engineering design task, namely Failure Modes and Effects Analysis (FMEA). A concrete example illustrates the advantages of the general approach in the context of FMEA.

Reasons for beliefs in understanding: applications of non-monotonic dependencies to story processing

National Conference on Artificial Intelligence, Aug 22, 1983

Many of the inferences and decisions which contribute to understanding involve fallible assumptio... more Many of the inferences and decisions which contribute to understanding involve fallible assumptions. When these assumptions are undermined, computational models of comprehension should respond rationally. This paper crossbreeds AI research on problem solving and understanding to produce a hybrid model ("reasoned understanding"). In particular, the paper shows how nonmonotonic dependencies [Doyle791 enable a schemabased story processor to adjust to new information requiring the retraction of assumptions.

Coherence and abduction

Behavioral and Brain Sciences, Sep 1, 1989

LT Revisited: Experimental Results of Applying Explanation-Based Learning to the Logic of Principia Mathematica

Elsevier eBooks, 1987

This paper descibes an explanation-based learning (EBL) system based on a version of Newell, Shaw... more This paper descibes an explanation-based learning (EBL) system based on a version of Newell, Shaw, and Simon's LOGIC-THEORIST (LT). Results of applying this system to propositional calculus problems from Principia Mathematica are compared with results of applying several other versions of the same performance element to these problems. The primary goal of this study is to characterize and analyze differences between non-learning, rote learning (LT's original learning method), and EBL. Another aim is to provide a characterization of the performance of a simple problem solver in the context of the Principia problems, in the hope that these problems can be used as a benchmark for testing improved learning methods, just as problems like chess and the eight puzzle have been used as benchmarks in research on search methods.

Abduction and learning : case studies in diverse domains

This paper presents a knowledge-based learning method and reports on case stud ies in different d... more This paper presents a knowledge-based learning method and reports on case stud ies in different domains. The method integrates abduction and learning. Abduction provides an improved method for constructing explanations, in comparison with de ductive methods traditionally associated with explanation-based learning. The im provement enlarges the set of examples that can be explained so that one can leam from additional examples using explanation-based macro-learning techniques. Ab duction also provides a form of knowledge level learning. The importance of abductive learning is shown by case studies involving over a hundred examples taken from diverse domains requiring logical, physical, and psychological knowledge and reasoning. The case studies are relevant to a wide range of practical tasks including; natural language understanding and plan recognition; qualitative physical reasoning and postdiction; diagnosis and signal interpretation; and decision-making imder un certainty. The descriptions of the case studies show how to set up abduction engines in particular domains and how abduction solves particular tasks. They show how to provide and how to represent the relevant knowledge, including both domain knowl edge and meta-knowledge relevant to abduction and search control. The description of each case study includes an example, its explanation, and discussions of what is learned by macro-learning and by abductive inference.

A qualitative logic of decision

An important aspect of intelligent behavior is the ability to reason, make decisions, and act in ... more An important aspect of intelligent behavior is the ability to reason, make decisions, and act in spite of uncertainty. This paper presents a qualitative logic of decision that supports decision-making under uncertainty. To be specific, the paper presents a knowledge representation language based upon subjective Bayesian decision theory that aims to capture some aspects of common-sense reasoning associated with making decisions about actions. The language addresses the problem of describing justifications of rational choices in situations where the alternatives involve trading off potential losses and gains. The logic and an associated qualitative arithmetic are implemented in an efficient PROLOG program. Examples illustrate their use in several concrete decision-making situations.

Working Notes of the 1990 Spring Symposium on Automated Abduction

scribe techniques for .constructing models on an asneeded basis. [Levitt et al., 1989] discusses ... more scribe techniques for .constructing models on an asneeded basis. [Levitt et al., 1989] discusses incremental model evaluation and extension. This approach is being tested in a program called Wimp3, which works as follows: l. A parser reads one word of the English text. It produces statements which describe the words in the story and the syntactic relations between them. 2. The output of the parser is taken by the network construction component. This component contains rules for language abduction. It builds a net, or extends the current net if some input has already been received. 3. The resulting belief network is evaluated by a network-evaluation component. If certain conclusions are overwhelmingly favored, they may be accepted as true to simplify further computation. 4. Return to step 1. This work is more fully described in three other papers of the authors: [Charniak and Goldman, 1989b] and [Charniakand Goldman, 1989a] give a probabilistic account of the •problem of story understanding. This provides the mathematical foundation of this work. [Goldman and Charniak, 1989) gives a more detailed account of the program. Finally, [Carroll and Charniak, 1989) discusses a marker-passer which is used to control the search of the belief network construction rules. Two other papers, [Goldman and Charniak, 1988) and [Charniak and Goldman, 1988) describe the authors' earlier work in this area, an attempt to build a logical-probabilistic hybrid program for story understanding.

Theory formation by abduction : a case study based on the chemical revolution

Abduction is the process of constructing explanations. This chapter suggests that automated abduc... more Abduction is the process of constructing explanations. This chapter suggests that automated abduction is a key to advancing beyond the "routine theory revision" methods developed in early AI research towards automated reasoning systems capable of "world model revision"-dramatic changes in systems of beliefs such as occur in children's cognitive development and in scientific revolutions. The chapter describes a general approach to automating theory revision based upon computational methods for theory formation by abduction. The approach is based on the idea that, when an anomaly is encountered, the best course is often simply to suppress parts of the original theory thrown into question by the contradiction and to derive an explanation of the anomalous observation based on relatively solid, basic principles. This process of looking for explanations of unexpected new phenomena can lead by abductive inference to new hypotheses that can form crucial part~ of a revised theory. As an illustration, the chapter shows how some of Lavoisier's key insights during the Chemical Revolution can be viewed as examples of theory formation by abduction.

Constraints for predicate invention

This chapter describes an inductive learning method that derives logic programs and invents predi... more This chapter describes an inductive learning method that derives logic programs and invents predicates when needed. The basic idea is to form the least common anti-instance (LCA) of selected seed examples. If the LCA is too general it forms the starting poínt of a general-to-specific search which is guided by various constraints on argument dependencies and critica! terms. A distinguishing feature of the method is its ability to introduce new predicates. Predicate invention involves three steps. First, the need for a new predicate is discovered and the arguments of the new predicate are determíned using the same constraints that guide the search. In the second step, instances of the new predicate are abductively inferred. These instances form the input for the last step where the definition of the new predicate is induced by recursively applying the method again. We also outline how such a system could be more tightly integrated with an abductive learning system

Theory formation by abduction : a case study based on the chemical revolution

Author(s): O'Rorke, Paul; Morris, Steven; Schulenburg, David | Abstract: Abduction is the pro... more Author(s): O'Rorke, Paul; Morris, Steven; Schulenburg, David | Abstract: Abduction is the process of constructing explanations. This chapter suggests that automated abduction is a key to advancing beyond the "routine theory revision" methods developed in early AI research towards automated reasoning systems capable of "world model revision" - dramatic changes in systems of beliefs such as occur in children's cognitive development and in scientific revolutions. The chapter describes a general approach to automating theory revision based upon computational methods for theory formation by abduction. The approach is based on the idea that, when an anomaly is encountered, the best course is often simply to suppress parts of the original theory thrown into question by the contradiction and to derive an explanation of the anomalous observation based on relatively solid, basic principles. This process of looking for explanations of unexpected new phenomena can le...

Abduction and learning : case studies in diverse domains

Author(s): O'Rorke, Paul | Abstract: This paper presents a knowledge-based learning method an... more Author(s): O'Rorke, Paul | Abstract: This paper presents a knowledge-based learning method and reports on case studies in different domains. The method integrates abduction and learning. Abduction provides an improved method for constructing explanations, in comparison with deductive methods traditionally associated with explanation-based learning. The improvement enlarges the set of examples that can be explained so that one can learn from additional examples using explanation-based macro-learning techniques. Abduction also provides a form of knowledge level learning. The importance of abductive learning is shown by case studies involving over a hundred examples taken from diverse domains requiring logical, physical, and psychological knowledge and reasoning. The case studies are relevant to a wide range of pratical tasks including: natural langauge understanding and plan recognition; qualitative physical reasoning and postdiction; diagnosis and signal interpretation; and decis...

Focusing on the most important explanations: decision-theoretic horn abduction

Proceedings of the Twelfth National Conference on Artificial Intelligence, Oct 5, 1994

This paper describes a new method, called Decision-Theoretic Horn Abduction (DTHA), for generatin... more This paper describes a new method, called Decision-Theoretic Horn Abduction (DTHA), for generating and focusing on the most important explanations. A procedure is given that can be used iteratively to generate a sequence of explanations from the most to the least important. The new method considers both the likelihood and utility of partial explanations and is applicable to a wide range of tasks. This paper shows how it applies to an important engineering design task, namely Failure Modes and Effects Analysis (FMEA). A concrete example illustrates the advantages of the general approach in the context of FMEA.

Qualitative Decision Analysis: Comparing alternative approaches

Qualitative decision analysis (QDA) extends qualitative reasoning methods developed forreasoning ... more Qualitative decision analysis (QDA) extends qualitative reasoning methods developed forreasoning about physical systems and for reasoning about probability to include informationabout utilities and values. QDA aims to solve problems associated with applying fullyquantitative approaches to decision analysis to problems in artificial intelligence, includingproblems associated with knowledge acquisition, inference, and explanation. A major weaknessof qualitative methods is that they are...

Generalization for Explanation-Based Schema Acquisition

This paper is about explanation-based learn- ing for heuristic problem solvers which "build&... more This paper is about explanation-based learn- ing for heuristic problem solvers which "build" solutions using schemata (frames like scripts) as both "bricks" and "mortar". The heart of the paper is a description of a generalization method which is designed to extract as much information as pos- sible from examples of successful problem solving behavior. A' related generalizer, (less powerful but more efficient), has been implemented as part of an experimental apprentice.* I INTRODUCTION Many knowledge-based AI systems have used schemata (knowledge packets such as frames or scripts) as the basis of computational models of understanding (I), planning (2) or other problem solving (3), but very few of these systems have been capable of generating their own schemata. As a result, most schema-based systems have been unable to automatically profit from their experi- ences, so that the main way of improving their performance has been by laboriously hand-codi...

LT Revisited: Explanation-Based Learning and the Logic of Principia Mathematica

Machine Learning

This paper descibes an explanation-based learning (EBL) system based on a version of Newell, Shaw... more This paper descibes an explanation-based learning (EBL) system based on a version of Newell, Shaw, and Simon's LOGIC-THEORIST (LT). Results of applying this system to propositional calculus problems from Principia Mathematica are compared with results of applying several other versions of the same performance element to these problems. The primary goal of this study is to characterize and analyze differences between non-learning, rote learning (LT's original learning method), and EBL. Another aim is to provide a characterization of the performance of a simple problem solver in the context of the Principia problems, in the hope that these problems can be used as a benchmark for testing improved learning methods, just as problems like chess and the eight puzzle have been used as benchmarks in research on search methods.

LT Revisited: Experimental Results of Applying Explanation-Based Learning to the Logic of Principia Mathematica

Proceedings of the Fourth International Workshop on MACHINE LEARNING, 1987

This paper descibes an explanation-based learning (EBL) system based on a version of Newell, Shaw... more This paper descibes an explanation-based learning (EBL) system based on a version of Newell, Shaw, and Simon's LOGIC-THEORIST (LT). Results of applying this system to propositional calculus problems from Principia Mathematica are compared with results of applying several other versions of the same performance element to these problems. The primary goal of this study is to characterize and analyze differences between non-learning, rote learning (LT's original learning method), and EBL. Another aim is to provide a characterization of the performance of a simple problem solver in the context of the Principia problems, in the hope that these problems can be used as a benchmark for testing improved learning methods, just as problems like chess and the eight puzzle have been used as benchmarks in research on search methods.

Acquiring knowledge relevant to tasks involving decision making

IntroductionMany tasks involve decision making under uncertainty.For example, in diagnosis and tr... more IntroductionMany tasks involve decision making under uncertainty.For example, in diagnosis and troubleshooting, often onemust decide whether to treat a suspected disease or malfunctionor to perform a diagnostic test to reduce uncertaintyabout the cause of observed symptoms. In planning,one must decide how best to achieve desired goals.Selection tasks such as selecting the most appropriateapplicant for a job, or the best product

Abductive signal interpretation for nondestructive evaluation

NDT & E International, 1996

This paper reports on one aspect of a collaborative project exploring applications of AI to Nonde... more This paper reports on one aspect of a collaborative project exploring applications of AI to Nondestructive Evaluation (NDE). The goal of NDE is to determine whether parts under evaluation are "good" or "defective" without damaging the parts. Piezoelectric probes are often used to produce ultrasonic signals that indicate the existence of cracks and other defects inside solid materials. This paper shows how knowledge-intensive explanatory reasoning can be used to construct an interpretation of the signals, simultaneously classifying the material under evaluation.

ABDUCTION and EXPLANATION-BASED LEARNING: CASE STUDIES IN DIVERSE DOMAINS

Computational Intelligence, 1994

This paper presents a knowledge-based learning method and reports on case studies in different do... more This paper presents a knowledge-based learning method and reports on case studies in different domains. The method integrates abduction and explanation-based learning. Abduction provides an improved method for constructing explanation$. The improvement enlarges the set of examples that can be explained so that one can learn from additional examples using traditional explanation-based macro learning. Abduction also provides a form of knowledge level learning. Descriptions of case studies show how to set up abduction engines for tasks in particular domains. The case studies involve over a hundred examples taken from diverse domains requiring logical, physical, and psychological knowledge and reasoning. The case studies are relevant to a wide range of practical tasks including natural language understanding and plan recognition; qualitative physical reasoning and postdiction; diagnosis and signal interpretation; and decision making under uncertainty. The descriptions of the case studies include an example, its explanation, and discussions of what is learned by macro-learning and by abductive inference. The paper discusses how to provide and represent the domain knowledge and meta-knowledge needed for abduction and search control. The main conclusion is that abductive inference is important for learning. Abduction and macro-learning are complementary and synergistic.