Model Mining

Model Mining

For programs, common behavior is often correct behavior. We have devised three approaches that capture behavior exhibited by programs written in Java. The results are abstract models that we use for program understanding and debugging.

Cebit 2010

• Pachika fixes programs by comparing program runs using object behavior models. See the press release.
• Javalanche assesses the quality of a test suite. See the press release.
• Tikanga mines the wisdom of 200 million lines of linux code. Check your own code at checkmycode.org or take a look at the press release.
• Lamarck mines evolution patterns using object usage models.

Projects

• ADABU

  Uses dynamic analysis to mine object behavior models that represent the way objects' state changes when they are used as targets of method calls.

• AMPLE

  Uses differences between program behavior exhibited in passing and failing runs to find classes that are likely defective.

• JADET

  Uses static analysis to mine object usage models that represent the way objects are being used in a program.

• API Birthmark

  Uses dynamic analysis to extract birthmarks, that capture instrinsic properties of a program, from program runs.

• PACHIKA

  Uses object behavior models to synthesize fixes.

• TAUTOKO

  Test case generation for typestate mining.

• TIKANGA

  Combines traditional static analysis with model checking to come up with CTL specifications for methods.

Papers

• Andrzej Wasylkowski, Andreas Zeller: Mining temporal specifications from object usage. 24th IEEE/ACM International Conference on Automated Software Engineering, Auckland, New Zealand, November 2009.
• Valentin Dallmeier, Andreas Zeller, Bertrand Meyer: Generating Fixes from Object Behavior Anomalies 2009.[PDF]
• Andrzej Wasylkowski, Andreas Zeller: Mining Operational Preconditions. Technical Report (updated), 2008.[PDF]
• Andrzej Wasylkowski, Andreas Zeller, Christian Lindig: Detecting Object Usage Anomalies. 6th joint meeting of the European Software Engineering Conference and the ACM SIGSOFT Symposium on the Foundations of Software Engineering (ESEC/FSE 2007), Cavtat near Dubrovnik, Croatia, September 2007. [DOI]
• Andrzej Wasylkowski: Mining Object Usage Models (Doctoral Symposium). 29th International Conference on Software Engineering (ICSE 2007), Minneapolis, MN, USA, May 2007. [DOI]
• Valentin Dallmeier, Christian Lindig, Andrzej Wasylkowski, Andreas Zeller: Mining Object Behavior with ADABU. Fourth International Workshop on Dynamic Analysis (WODA 2006), Neelam Gupta, Andy Podgurski (Eds.), Shanghai, China, May 2006.
• Valentin Dallmeier, Christian Lindig, Andreas Zeller: Lightweight Bug Localization with AMPLE (Demo Paper). Sixth International Symposium on Automated and Analysis-Driven Debugging, Jong Deok Choi and Raimondas Lencevicius (Eds.), pages 99-103, ACM Press, Monterey, CA, USA, September 2005.
• Valentin Dallmeier, Christian Lindig, Andreas Zeller: Evaluating a Lightweight Defect Localization Tool. PLDI Workshop on the Evaluation of Software Defect Detection Tools, 99--103, Chicago, June 2005. Proceedings were only distributed at the workshop.
• Valentin Dallmeier, Christian Lindig, Andreas Zeller: Lightweight Defect Localization for Java. Proceedings of 19th European Conference on Object-Oriented Programming, ECOOP 2005, Andrew P. Black (Ed.), pages 528--550, Lecture Notes in Computer Science 3586, Springer, Glasgow, UK, July 2005.
• Christian Lindig, Valentin Dallmeier, Andreas Zeller: 7th Workshop Software Reengineering: Lightweight Control-Flow Abstraction. Softwaretechnik-Trends, 41-42, (25) 2, Gesellschaft für Informatik, May 2005.

Contact us

Below you can find links to the people working on model mining. Please contact us for questions, rants, and raves.

• Valentin Dallmeier
• Christian Lindig
• Andrzej Wasylkowski
• Andreas Zeller