Reduxexp: An Open-source Justification-based Explanation Support Server

Post on 15-Dec-2014

569 views 0 download

Tags:

description

Trust in any decision depends on many factors. One way of improving trust is by explaining why a certain decision has been made and how certain actions in the context of a decision have been performed. In order to provide such explanations information systems need a mechanism for recording decisions and actions. The open-source server Reduxexp is such a system. This paper describes the architecture of Reduxexp and first experiences with the system.

transcript

Reduxexp: An Open-source Justification-based Explanation Support Server

Thomas Roth-Berghofer & Florian MittagSenior researcher, trb@dfki.de German Research Centre for Artificial Intelligence DFKI GmbH

CAMBRIDGE, UK, 11 DEC 2008

Samstag, 18. Juli 2009

What are explanations?

Samstag, 18. Juli 2009

What are explanations?

Samstag, 18. Juli 2009

What are explanations?Explanations are answers to questions.

Samstag, 18. Juli 2009

Explanations are a kind of information.

Samstag, 18. Juli 2009

Concepts connected with the notion of information

THE COACH TO STANSTED AIRPORT LEAVES

CAMBRIDGE AT 17:30 H.

Samstag, 18. Juli 2009

• Truth

Concepts connected with the notion of information

THE COACH TO STANSTED AIRPORT LEAVES

CAMBRIDGE AT 17:30 H.

Samstag, 18. Juli 2009

• Truth

• Understanding

Concepts connected with the notion of information

THE COACH TO STANSTED AIRPORT LEAVES

CAMBRIDGE AT 17:30 H.

Samstag, 18. Juli 2009

• Truth

• Understanding

• Utility

Concepts connected with the notion of information

THE COACH TO STANSTED AIRPORT LEAVES

CAMBRIDGE AT 17:30 H.

Samstag, 18. Juli 2009

• Truth

• Understanding

• Utility

• Storage and retrieval

Concepts connected with the notion of information

THE COACH TO STANSTED AIRPORT LEAVES

CAMBRIDGE AT 17:30 H.

Samstag, 18. Juli 2009

• Truth

• Understanding

• Utility

• Storage and retrieval

• Information generation

Concepts connected with the notion of information

THE COACH TO STANSTED AIRPORT LEAVES

CAMBRIDGE AT 17:30 H.

Samstag, 18. Juli 2009

• Truth

• Understanding

• Utility

• Storage and retrieval

• Information generation

Concepts connected with the notion of information

THE COACH TO STANSTED AIRPORT LEAVES

CAMBRIDGE AT 17:30 H.

Samstag, 18. Juli 2009

Basic explanation scenario

Samstag, 18. Juli 2009

Knowledge sources

Samstag, 18. Juli 2009

Problem solving

knowledge

Knowledge sources

Samstag, 18. Juli 2009

Problem solving

knowledge

Explanation

knowledge

Knowledge sources

Samstag, 18. Juli 2009

Problem solving

knowledge

Explanation

knowledge

Knowledge sources

Samstag, 18. Juli 2009

Example: course assignment

• Professors Descartes and Kant

• Descartes more senior than Kant

• Heuristics:• Try to assign a course to the professor who wants

to teach that course.

• Assign courses to senior professors first.

• Preferences:• Descartes wants to teach course PHL305.

• Kant also wants to teach course PHL305.

• Kant would also teach course PHL380.Immanuel Kant

1724–1804

René Descartes1596–1650

Samstag, 18. Juli 2009

Example: course assignment

• Decisions:• Descartes will teach PHL305.

• Kant will teach PHL380.

• What happens if Descartes could not teach PHL305?

Immanuel Kant 1724–1804

René Descartes1596–1650

Samstag, 18. Juli 2009

Justification-based Truth Maintenance Systems

• Network of nodes and justifications, which are labelled IN or OUT:• Node is labelled IN <=> at least one of its supporting

justifications is labelled IN

• Justification is labelled IN <=> each element in its IN-list is labelled IN and each element in its OUT-list i labelled OUT

Samstag, 18. Juli 2009

REDUX

Samstag, 18. Juli 2009

REDUX

G1

Samstag, 18. Juli 2009

REDUX

G1

O1

O2

Samstag, 18. Juli 2009

REDUX

G1Conflict

set

O1

O2

Samstag, 18. Juli 2009

REDUX

G1Conflict

set

Decision

O1

O2

Samstag, 18. Juli 2009

REDUX

G1Conflict

set

Decision

Reduction

O1

O2

GnG2

Samstag, 18. Juli 2009

REDUX

G1Conflict

set

Decision

Reduction

A1, A2, …

O1

O2

GnG2

Samstag, 18. Juli 2009

REDUX

G1Conflict

set

Decision

Reduction

A1, A2, …

O1

O2

GnG2

: Goal reduction

Samstag, 18. Juli 2009

Reduxexp Architecture

• Provenance• who decided

• and when

• Persistence• serialisation of network

into RDF/XML

Samstag, 18. Juli 2009

Example:course assignment

Samstag, 18. Juli 2009

Example:course assignment

Samstag, 18. Juli 2009

Example:course assignment

Samstag, 18. Juli 2009

Example:course assignment

Samstag, 18. Juli 2009

Example:course assignment

• Kant enquires about his assignment:

Samstag, 18. Juli 2009

Example:course assignment

• Kant enquires about his assignment:

FOR EACH assignment IN redux.getAllAssignments() { IF assignment.getName() == "Kant teaches PHL305" RETURN assignment; } IF assignment != null PRINT assignment.isValid();

Samstag, 18. Juli 2009

Example:course assignment

• Kant: “Why didn’t I get to teach PHL305?”decisions[] = redux.getDecisionsForAssignment(assignment);

Samstag, 18. Juli 2009

Example:course assignment

• Kant: “Why didn’t I get to teach PHL305?”decisions[] = redux.getDecisionsForAssignment(assignment);

rejections[] = decisions[0].getRejectionReasons();

Samstag, 18. Juli 2009

Example:course assignment

• Kant: “Why didn’t I get to teach PHL305?”decisions[] = redux.getDecisionsForAssignment(assignment);

rejections[] = decisions[0].getRejectionReasons();

ReduxReason rejection = rejections[0]; PRINT "Assignments: "; FOR EACH a IN rejection.getConditionAssignments { PRINT a.getName() + " by " + a.getCreator(); } PRINT "Facts: "; FOR EACH f IN rejection.getConditionFacts { PRINT f.getName() + " by " + f.getCreator(); } PRINT "Rejected decisions: "; FOR EACH d IN rejection.getConditionDecisions { PRINT d.getName() + " by " + d.getCreator(); }

Samstag, 18. Juli 2009

Example:course assignment

• Output:Assignments: "Descartes teaches PHL305" by ’Administrator’ Facts: "At most one teacher per course" by ’Administrator’ Rejected decisions: none

Samstag, 18. Juli 2009

Example:course assignment

• Output:Assignments: "Descartes teaches PHL305" by ’Administrator’ Facts: "At most one teacher per course" by ’Administrator’ Rejected decisions: none

Note:

• Reduxexp does not generate explanations, it supports the process by providing the necessary information

Samstag, 18. Juli 2009

Example:course assignment

• Output:Assignments: "Descartes teaches PHL305" by ’Administrator’ Facts: "At most one teacher per course" by ’Administrator’ Rejected decisions: none

Note:

• Reduxexp does not generate explanations, it supports the process by providing the necessary information

• Client application supplies semantics.

Samstag, 18. Juli 2009

Take home messages

Samstag, 18. Juli 2009

• Explanations are important for increasing the trustworthiness of decision support systems.

Take home messages

Samstag, 18. Juli 2009

• Explanations are important for increasing the trustworthiness of decision support systems.

• Explanations are part of communication scenario.

Take home messages

Explainer

Originator

User

Samstag, 18. Juli 2009

• Explanations are important for increasing the trustworthiness of decision support systems.

• Explanations are part of communication scenario.

• Explanation provision is only possible if the respective information is available.

Take home messages

Explainer

Originator

User

Samstag, 18. Juli 2009

• Explanations are important for increasing the trustworthiness of decision support systems.

• Explanations are part of communication scenario.

• Explanation provision is only possible if the respective information is available.

• Reduxexp is a server for logging / recording decisions.

Take home messages

Explainer

Originator

User

MORE INFORMATION AND DOWNLOAD: HTTP://REDUXEXP.OPENDFKI.DE

Samstag, 18. Juli 2009

Thomas Roth-Berghofer & Florian MittagSenior researcher, trb@dfki.de German Research Centre for Artificial Intelligence DFKI GmbH

Thank you!

Reduxexp: An Open-source Justification-based Explanation Support Server

CAMBRIDGE, UK, 11 DEC 2008

Samstag, 18. Juli 2009

Invitation to participate

• ExaCt mailing list:http://groups.yahoo.com/group/explanation-research/

• ExaCt 2009 @ IJCAI 2009 http://exact2009.workshop.hm

Samstag, 18. Juli 2009

Publications of ExaCt research group

2008

[Bahls, 2008] Bahls, D. (2008). Explanation support for the case-based reasoning tool myCBR. Project thesis, University of Kaiserslautern.

[Forcher et al., 2008] Forcher, B., Adrian, B., and Roth-Berghofer, T. (2008). Explanations in the information extraction system iDocument. Künstliche Intelligenz, 22(2).

[Mittag, 2008] Mittag, F. (2008). ReduxExp: A justification-based explanation-support server. Project thesis, University of Kaiserslautern.

[Roth-Berghofer and Bahls, 2008] Roth-Berghofer, T. R. and Bahls, D. (2008). Code tagging and retrieval with myCBR. In Petridis, M., Coenen, F., and Bramer, M., editors, Research and Development in Intelligent Systems XXV, London, UK. Springer Verlag.

[Roth-Berghofer and Mittag, 2008] Roth-Berghofer, T. R. and Mittag, F. (2008). ReduxExp: A justification-based explanation-support server. Proceedings of AI-2008. the twenty-eighth SGAI international conference on artificial intelligence. In Petridis, M., Coenen, F., and Bramer, M., editors, Research and Development in Intelligent Systems XXV, London, UK. Springer Verlag.

[Roth-Berghofer and Richter, 2008a] Roth-Berghofer, T. R. and Richter, M. M., editors (2008a). Künstliche Intelligenz—Topic: Explanation, volume 22, Bremen. BöttcherIT Verlag.

[Roth-Berghofer and Richter, 2008b] Roth-Berghofer, T. R. and Richter, M. M. (2008b). On explanation. Künstliche Intelligenz, 22(2):5–7.

2007

[Bahls and Roth-Berghofer, 2007] Bahls, D. and Roth-Berghofer, T. (2007). Explanation support for the case-based reasoning tool myCBR. In Proceedings of the Twenty-Second AAAI Conference on Artificial Intelligence. July 22–26, 2007, Vancouver, British Columbia, Canada., pages 1844–1845. The AAAI Press, Menlo Park, California.

[Eppert, 2007] Eppert, M. (2007). Generating provenance explanations for the gnowsis rebirth machine - a first pass. Pro ject thesis, University of Kaiserslautern.

2007 (contd.)

[Roth-Berghofer et al., 2007] Roth-Berghofer, T. R., Schulz, S., and Leake, D. B., editors (2007). Proceedings of the AAAI-07 workshop on Explanation-aware Computing ExaCt2007. AAAI Press. Technical Report WS-07-06.

2006

[Richter et al., 2006] Richter, M. M., Roth-Berghofer, T., and Schulz, S., editors (2006). Explanation-aware Computing, volume 25. SAP - Slovak Academic Press Ltd., Bratislava.

2005

[Roth-Berghofer et al., 2005a] Roth-Berghofer, T., Cassens, J., and Sørmo, F. (2005a). Goals and kinds of explanations in case-based reasoning. In Althoff, K.-D., Dengel, A., Bergmann, R., Nick, M., and Roth-Berghofer, T., editors, WM 2005: Professional Knowledge Management, pages 264–268, Kaiserslautern, Germany. DFKI GmbH.

[Roth-Berghofer and Cassens, 2005] Roth-Berghofer, T. R. and Cassens, J. (2005). Mapping goals and kinds of explanations to the knowledge containers of case-based reasoning systems. In Muñoz-Avila, H. and Ricci, F., editors, Case-Based Reasoning Research and Developmen, pages 451–464, Heidelberg. Springer Verlag.

[Roth-Berghofer et al., 2005b] Roth-Berghofer, T. R., Schulz, S., and Woody, A., editors (2005b). Proceedings of the AAAI Fal l Symposium on Explanation-aware Computing ExaCt2005. AAAI Press. Technical Report FS-05-04.

2004

[Roth-Berghofer, 2004] Roth-Berghofer, T. R. (2004). Explanations and Case-Based Reasoning: Foundational issues. In Funk, P. and González-Calero, P. A., editors, Advances in Case-Based Reasoning, pages 389–403. Springer-Verlag.

Samstag, 18. Juli 2009