Academia.eduAcademia.edu

Outline

Title
Abstract
Introduction
Overview
Previous Work on Learning at the Yale Ai Lab
An Algorithm for Case-Based Explanation
Substitute Related Action
Conclusion
References
All Topics
Psychology
Cognitive Science

Creativity and learning in a case-based explainer

Profile image of David LeakeDavid Leake

1989, Artificial Intelligence

https://doi.org/10.1016/0004-3702(89)90053-2
visibility

…

description

33 pages

link

1 file

Abstract

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 explain them by remembering old explanations, and adapting them to fit. We believe that this case-based approach to explanation holds promise for use in AI systems, both for routine explanation and to creatively explain situations quite unlike what the system has encountered before. Building new explanations from old ones relies on having explanations available in memory. We describe explanation patterns (XPs), knowledge structures that package the reasoning underlying explanations. Using the SWALE system as a base, we discuss the retrieval and modification process, and the criteria used when deciding which explanation to accept. We also discuss issues in learning XPs: what generalization strategies are appropriate for real-world explanations, and which indexing strategies are appropriate for XPs. SWALE' s explanations allow it to understand nonstandard stories, and the XPs it learns increase its efficiency in dealing with similar anomalies in the future.

Related papers

Explanation In Case-Based Reasoning - Perspectives and Goals
Jörg Cassens

Artificial Intelligence Review, 2005

View PDFchevron_right
Explanation-driven retrieval, reuse, and learning of cases
Agnar Aamodt

1993

A method for integrated case-based and generalization-based reasoning and learning is described. The primary role of general domain knowledge is to provide explanatory support for the case-based processes. A general explanation engine -the ACTIVATE-EXPLAIN-FOCUS cycle -utilizes a presumably rich, multirelational knowledge model in producing context-dependent explanations.

View PDFchevron_right
Explanation Goals In Case-Based Reasoning
Jörg Cassens

Proceedings of the ECCBR, 2004

View PDFchevron_right
Gaining insight through case-based explanation
Padraig Cunningham

2009

Abstract Traditional explanation strategies in machine learning have been dominated by rule and decision tree based approaches. Case-based explanations represent an alternative approach which has inherent advantages in terms of transparency and user acceptability. Case-based explanations are based on a strategy of presenting similar past examples in support of and as justification for recommendations made.

View PDFchevron_right
A top-level model of case-based argumentation for explanation: Formalisation and experiments
Rosa Ratsma

Argument & Computation, 2021

This paper proposes a formal top-level model of explaining the outputs of machine-learning-based decision-making applications and evaluates it experimentally with three data sets. The model draws on AI & law research on argumentation with cases, which models how lawyers draw analogies to past cases and discuss their relevant similarities and differences in terms of relevant factors and dimensions in the problem domain. A case-based approach is natural since the input data of machine-learning applications can be seen as cases. While the approach is motivated by legal decision making, it also applies to other kinds of decision making, such as commercial decisions about loan applications or employee hiring, as long as the outcome is binary and the input conforms to this paper’s factor- or dimension format. The model is top-level in that it can be extended with more refined accounts of similarities and differences between cases. It is shown to overcome several limitations of similar arg...

View PDFchevron_right
A review of explanation and explanation in case-based reasoning
Padraig Cunningham

2003

With the arrival of more powerful computers and improved algorithms in machine learning, the availability of computer-based Knowledge Based Systems (KBSs) has increased rapidly. There is a wide range of KBSs available in many different types of domains, including medicine, finance, industry and technical diagnosis (Armengol et al, 2000; Ong et al, 1997; Rowe & Wright, 1993; Mark et al, 1996; Ye, 1995). Sales of KBS development tools are growing at a rate of about 16% per annum since 1988 (Durkin, 1996).

View PDFchevron_right
Explainable AI through the Learning of Arguments
Nico Roos

arXiv (Cornell University), 2022

Learning arguments is highly relevant to the field of explainable artificial intelligence. It is a family of symbolic machine learning techniques that is particularly human-interpretable. These techniques learn a set of arguments as an intermediate representation. Arguments are small rules with exceptions that can be chained to larger arguments for making predictions or decisions. We investigate the learning of arguments, specifically the learning of arguments from a 'case model' proposed by Verheij [32]. The case model in Verheij's approach are cases or scenarios in a legal setting. The number of cases in a case model are relatively low. Here, we investigate whether Verheij's approach can be used for learning arguments from other types of data sets with a much larger number of instances. We compare the learning of arguments from a case model with the HeRO algorithm [15] and learning a decision tree.

View PDFchevron_right
Goals and Kinds of Explanations In Case-Based Reasoning
Jörg Cassens

WM, 2005

View PDFchevron_right
An evaluation of the usefulness of case-based explanation
Padraig Cunningham

2003

One of the perceived benefits of Case-Based Reasoning (CBR) is the potential to use retrieved cases to explain predictions. Surprisingly, this aspect of CBR has not been much researched. There has been some early work on knowledge-intensive approaches to CBR where the cases contain explanation patterns (eg SWALE). However, a more knowledge-light approach where the case similarity is the basis for explanation has received little attention.

View PDFchevron_right
Scaling Up Explanation Generation: Large-Scale Knowledge irical St ies
James Lester

Abstract To explain complex phenomena, an explanation system must be able to select information from a formal representation of domain knowledge, organize the selected information into multisentential discourse plans, and realize the discourse plans in text. Although recent years have witnessed significant progress in the development of sophisticated computational mechanisms for explanation, empirical results have been limited.

View PDFchevron_right

References (32)

  1. Carbonell, J.G., Learning by analogy: Formulating and generalizing plans from past ex- perience, in: R.S. Michalski, J.G. CarboneU and T.M. Mitchell (Eds.), Machine Learning: An Artificial Intelligence Approach (Tioga, Palo Alto, CA, 1983).
  2. Carbonell, J.G., Derivational analogy: A theory of reconstructive problem solving and expertise acquisition, in: R.S. Michalski, J.G. Carbonell and T.M. Mitchell (Eds.), Machine Learning: An Artificial Intelligence Approach 2 (Morgan Kaufmann, Los Altos, CA, 1986) 371-392.
  3. Cullingford, R., Script application: Computer understanding of newspaper stories, Ph.D. Thesis, Tech. Rept. 116, Yale University, New Haven, CT (1978).
  4. DeJong, J. and Mooney, R., Explanation-based learning: An alternative view, Mach. Learn- ing 1 (1986) 145-176.
  5. Dietterich, T. and Michalski, R., A comparative review of selected methods for learning from examples, in: R.S. Michalski, J.G. Carbonell and T.M. Mitchell (Eds.), Machine Learning: An Artificial Intelligence Approach (Tioga, Paio Alto, CA, 1983) 41-81.
  6. Gentner, D., Structure-mapping: A theoretical framework for analogy, Cognitive Psychol. 7 (1983) 155-170.
  7. Gentner, D. and Landers, R., Analogical reminding: A good match is hard to find, in: Proceedings IEEE 1985 International Conference on Systems, Man and Cybernetics (1985) 607ff.
  8. Gick, M.L. and Holyoak, K.J., Analogical problem solving, Cognitive Psychol. 12 (1980) 306-355.
  9. Gick, M.L. and Holyoak, K.J., Schema induction and analogical transfer, Cognitive Psychol. 15 (1983) 1-38.
  10. Hammond, K.J., Case-based planning: An integrated theory of planning, learning and memory, Ph.D. Thesis, Tech. Rept. 488, Yale University, New Haven, CT (1986).
  11. Holyoak, K.J., The pragmatics of analogical transfer, in: G.H. Bower (Ed.), The Psychology of Learning and Motivation (Academic Press, Orlando, FL, 1985).
  12. Kass, A., Modifying explanations to understand stories, in: Proceedings Eighth Annual Conference of the Cognitive Science Society, Amherst, MA (1986).
  13. Kass, A. and Leake, D.B., Types of explanations, Tech. Rept. 523, Yale University, Department of Computer Science, New Haven, CT (1987).
  14. Kedar-Cabelli, S.T., Purpose-directed analogy, in: Proceedings Seventh Annual Conference of the Cognitive Science Society, Irvine, CA (1985).
  15. Koiodner, J.L., Retrieval and organizational strategies in conceptual memory: A computer model, Ph.D. Thesis, Tech. Rept. 187, Yale University, New Haven, CT (1980).
  16. Kolodner, J., Simpson, R. and Sycara, K., A process model of case-based reasoning in problem solving, in: A. Joshi (Ed.), Proceedings IJCA1-85, Los Angeles, CA (1985) 284-290.
  17. Leake, D.B. and Owens, C., Organizing memory for explanation, in: Proceedings Eighth Annual Conference of the Cognitive Science Society, Amherst, MA (1986).
  18. Lebowitz, M., Generalization and memory in an integrated understanding system, Ph.D. Thesis, Tech. Rept. 186, Yale University, New Haven, CT (1980).
  19. Lebowitz, M., Integrated learning: Controlling explanation, Cognitive Sci. 10 (1986) 219-240.
  20. Mitchell, T.M., Generalization as search, Artificial Intelligence 18 (1982) 203-226.
  21. Mitchell, T.M., Keller, R.M. and Kedar-Cabelli, S.T., Explanation-based generalization: A unifying view, Mach, Learning 1 (1986) 47-80.
  22. Rieger, C., Conceptual memory and inference, in: R.C. Schank (Ed.), Conceptual Information Processing (North-Holland, Amsterdam, 1975).
  23. Riesbeck, C.K., Failure-driven reminding for incremental learning, in: Proceedings IJCAI-81, Vancouver, BC (1981) 115-120.
  24. Schank, R.C., Dynamic Memory: A Theory of Learning in Computers and People (Cambridge University Press, Cambridge, 1982).
  25. Schank, R.C., The current state of AI: One man's opinion, AI Mag. 4 (1) (1983) 3-8.
  26. Schank, R.C., Explanation: A first pass, Tech. Rept. 330, Yale University, Department of Computer Science, New Haven, CT (1984).
  27. Schank, R.C., Explanation Patterns: Understanding Mechanically and Creatively (Erlbaum, Hillsdale, NJ, 1986).
  28. Schank, R.C. and Abelson, R., Scripts, Plans, Goals and Understanding (Erlbaum, Hillsdale, NJ, 1977).
  29. Schank, R.C., Collins, G. and Hunter, L., Transcending inductive category formation in learning, Behav. Brain Sci. 9 (4) (1986).
  30. Schank, R.C. and Leake, D.B., Computer understanding and creativity, in: H.-J. Kugler (Ed.), Information Processing 86 (Elsevier Science Publishers B.V. (North-Holland), Am- sterdam, 1986) 335-341.
  31. Simpson, R.L., A computer model of case-based reasoning in problem-solving: An investiga- tion in the domain of dispute mediation, Ph.D. Thesis, School of Information and Computer Science, Georgia Institute of Technology, Atlanta, GA (1985).
  32. Sycara, E.P., Resolving adversarial conflicts: An approach integrating case-based and analytic methods, Ph.D. Thesis, School of Information and Computer Science, Georgia Institute of Technology, Atlanta, GA (1987).

Related papers

Adaptive similarity assessment for case-based explanation
David Leake

1995

Guiding the generation of abductive explanations is a di cult problem. Applying casebased reasoning to abductive explanation generation|generating new explanations by retrieving and adapting explanations for prior episodes|o ers the bene t of re-using successful explanatory reasoning but raises new issues concerning how to perform similarity assessment to judge the relevance of prior explanations to new situations. Similarity assessment a ects two points in the case-based explanation process: deciding which explanations to retrieve and evaluating the retrieved candidates. We address the problem of identifying similar explanations to retrieve by basing that similarity assessment on a categorization of anomaly types. We show that the problem of evaluating retrieved candidate explanations is often impeded by incomplete information about the situation to be explained, and address that problem with a novel similarity assessment method which we call constructive similarity assessment. Constructive similarity assessment contrasts with traditional \feature-mapping" similarity assessment methods by using the contents of memory to hypothesize important features in the new situation, and in using a pragmatic criterion|the system's ability to adapt features of the old case into features that apply in the new circumstances|as the basis for comparing features. Thus constructive similarity assessment does not merely compare new cases to old; instead, based on adaptation of prior cases in memory, it addresses the problem of incomplete input cases by building up and reasoning about augmented descriptions of those cases.

View PDFchevron_right
Mapping Goals and Kinds of Explanations to the Knowledge Containers of Case-Based Reasoning Systems
Jörg Cassens

Case-Based Reasoning Research and …, 2005

View PDFchevron_right
Compositional Adaptation of Explanations in Textual Case-Based Reasoning
pinar ozturk

When problem solving systems are deployed in real life, it is usually not enough to provide only a solution without any explanation. Users need an explanation in order to trust the system's decisions. At the same time, explanations may also function internally in the system's own reasoning process. One way to come up with an explanation for a new problem is to adapt an explanation from a similar problem encountered earlier, which is the idea behind the case-based explanation approach introduced by [29]. The original approach relies on manual construction of cases with explanations, which is difficult to scale up. In earlier work, therefore, we developed a system for automatic acquisition of cases with explanations from textual reports, including retrieval and adaptation of such cases [33,32]. In this paper, we improve the adaptation method by combining explanations from more than one case, which we call com-positional adaptation. The method is evaluated on an incident analysis task where the goal is to identify the root causes of a transportation incident , explaining it in terms of the information contained in the incident description. The evaluation results show that the proposed approach increases both the recall and the precision of the system.

View PDFchevron_right
Let Me Explain: Adaptation of Explanations Extracted from Incident Reports
pinar ozturk

Ai Communications , 2017

In order for us to trust artificial intelligence systems, they need to be able to explain their decisions. This article investigates a case-based reasoning (CBR) approach for generating explanations where a new event is explained by combining and modifying explanations of multiple previous events. This approach is implemented in a CBR system for incident analysis where the goal is to identify causes of transportation incidents. The system generates explanations that connect the observed events to the root causes of the incident through intermediate states and events. Explanations for past incidents are automatically extracted from textual incident reports using natural language processing, thus avoiding the manual effort of constructing explanations. The system is evaluated on incident reports from the Transportation Safety Board of Canada, supporting the hypothesis that adapting explanations of multiple previous events rather than a single event results in a more comprehensive explanation of a new event.

View PDFchevron_right
Explanation-driven case-based reasoning
Agnar Aamodt

1994

Problem solving in weak theory domains should compensate for the lack of strong theories by combining the various other knowledge types involved. Such methods should be able to effectively combine general domain knowledge with specific case knowledge. A method is described that utilises a presumably extensive and dense model of general domain knowledge as explanatory support for case-based problem solving and learning.

View PDFchevron_right

Related topics

  • Cognitive Science
  • Computer Science
  • Artificial Intelligence
  • Creativity

    • Cited by

    Fuzzy exemplar-based inference system for flood forecasting
    Li-Chiu Chang

    Water Resources Research, 2005

    1] Fuzzy inference systems have been successfully applied in numerous fields since they can effectively model human knowledge and adaptively make decision processes. In this paper we present an innovative fuzzy exemplar-based inference system (FEIS) for flood forecasting. The FEIS is based on a fuzzy inference system, with its clustering ability enhanced through the Exemplar-Aided Constructor of Hyper-rectangles algorithm, which can effectively simulate human intelligence by learning from experience. The FEIS exhibits three important properties: knowledge extraction from numerical data, knowledge (rule) modeling, and fuzzy reasoning processes. The proposed model is employed to predict streamflow 1 hour ahead during flood events in the Lan-Yang River, Taiwan. For the purpose of comparison the back propagation neural network (BPNN) is also performed. The results show that the FEIS model performs better than the BPNN. The FEIS provides a great learning ability, robustness, and high predictive accuracy for flood forecasting.

    View PDFchevron_right
    MEMORY ORGANIZATION AS THE MISSING LINK BETWEEN CASE-BASED REASONING AND INFORMATION RETRIEVAL IN BIOMEDICINE
    piao piao

    Computational Intelligence, 2006

    Mémoire proposes a general framework for reasoning from cases in biology and medicine. Part of this project is to propose a memory organization capable of handling large cases and case bases as occur in biomedical domains. This article presents the essential principles for an efficient memory organization based on pertinent work in information retrieval. Information retrieval systems have been able to scale up to terabyte of data taking advantage of large databases research to build Internet search engines. They search for pertinent documents to answer a query using term-based ranking and/or global ranking schemes. Similarly, CBR systems search for pertinent cases using a scoring function for ranking the cases. Mémoire proposes a memory organization based on inverted indexes, which may be powered by databases to search and rank efficiently through large case bases. It can be seen as a first step toward largescale CBR systems, and in addition provides a framework for tight cooperation between CBR and IR.

    View PDFchevron_right
    PROTOTYPICAL CASES FOR RETRIEVAL, REUSE, AND KNOWLEDGE MAINTENANCE IN BIOMEDICAL CASE-BASED REASONING
    piao piao

    Computational Intelligence, 2009

    Representing biomedical knowledge is an essential task in biomedical informatics intelligent systems. Casebased reasoning (CBR) holds the promise to represent contextual knowledge in a way that was not possible before with traditional knowledge-based methods. One main issue in biomedical CBR is dealing with the rate of generation of new knowledge in biomedical fields, which often makes the content of a case base partially obsolete. This article proposes to make use of the concept of prototypical case to ensure that a CBR system would keep update with current research advances in the biomedical field. Prototypical cases have served various purposes in biomedical CBR systems, among which to organize and structure the memory, to guide the retrieval as well as the reuse of cases, and to serve as bootstrapping a CBR system memory when real cases are not available in sufficient quantity and/or quality. This paper emphasizes the different roles prototypical cases can play in CBR systems, and presents knowledge maintenance as a very important novel role for these prototypical cases.

    View PDFchevron_right
    Problem solving for redesign
    Anita Pos, J. Akkermans, Remco Straatman

    Lecture Notes in Computer Science, 1997

    A knowledge-level analysis of complex tasks like diagnosis and design can give us a better understanding of these tasks in terms of the goals they aim to achieve and the different ways to achieve these goals. In this paper we present a knowledge-level analysis of redesign. Redesign is viewed as a family of methods based on some common principles, and a number of dimensions along which redesign problem solving methods can vary are distinguished. By examining the problem-solving behavior of a number of existing redesign systems and approaches, we came up with a collection of problem-solving methods for redesign and developed a task-method structure for redesign. In constructing a system for redesign a large number of knowledge-related choices and decisions are made. In order to describe all relevant choices in redesign problem solving, we have to extend the current notion of possible relations between tasks and methods in a PSM architecture. The realization of a task by a PSM, and the decomposition of a PSM into subtasks are the most common relations in a PSM architecture. However, we suggest to extend This work has been funded by NWO/SION within project 612-322-316, "Evolutionary design in knowledge-based systems" (the REVISE-project). Participants in the REVISE-project are: the TWIST group at the University of Twente, the SWI department of the University of Amsterdam, the AI department of the Vrije Universiteit van Amsterdam and the STEVIN group at the University of Twente. these relations with the notions of task refinement and method refinement. These notions represent intermediate decisions in a task-method structure, in which the competence of a task or method is refined without immediately paying attention to the operationalization in terms of subtasks. Explicit representation of this kind of intermediate decisions helps to make and represent decisions in a more piecemeal fashion.

    View PDFchevron_right
    Investigating Textual Case-Based XAI
    Rosina Weber

    Case-Based Reasoning Research and Development

    This paper demonstrates how case-based reasoning (CBR) can be used for an explainable artificial intelligence (XAI) approach to justify solutions produced by an opaque learning method (i.e., target method), particularly in the context of unstructured textual data. Our general hypothesis is twofold: 1) There exists patterns in the relationship between problems and solutions and there should be data or a body of knowledge that describes how problems and solutions relate; and 2) the identification, manipulation, and learning of such patterns through case features can help create and reuse explanations for solutions produced by the target method. When the target method relies on neural network architectures (e.g., deep learning), the resulting latent space (i.e., word embeddings) becomes useful for finding patterns and semantic relatedness in textual data. In the proposed approach, case problems are input-output pairs from the target method, and case solutions are explanations. We exemplify our approach by explaining recommended citations from Citeomatic  a multi-layer neural-network architecture from the Allen Institute for Artificial Intelligence. Citation analysis is the body of knowledge that describes how query documents (i.e., inputs) relate to recommended citations (i.e., outputs). We build cases and similarity assessment to learn features that represent patterns between problems and solutions that can lead to the reuse of corresponding explanations. The illustrative implementation we present becomes an explanation-augmented citation recommender that targets human-computer trust.

    View PDFchevron_right
    580 California St., Suite 400
    San Francisco, CA, 94104
    © 2025 Academia. All rights reserved