Learning to Adapt for Case-Based Design*

Nirmalie Wiratunga1, Susan Craw1, and Ray Rowe2

1School of Computing,
The Robert Gordon University,
Aberdeen AB25 1HG, Scotland, UK
{nw | s.craw}@scms.rgu.ac.uk
2 AstraZeneca
Macclesfield, Cheshire SK1O 2NA, England, UK




Abstract. Design is a complex open-ended task and it is unreasonable to expect a
case-base to contain representatives of all possible designs. Therefore, adaptation
is a desirable capability for case-based design systems, but acquiring adaptation
knowledge can involve significant effort. In this paper adaptation knowledge is
induced separately for different criteria associated with the retrieved solution, using 
knowledge sources implicit in the case-base. This provides a committee of
learners and their combined advice is better able to satisfy design constraints and
compatibility requirements compared to a single learner. The main emphasis of the
paper is to evaluate the impact of specific-to-general and general-to-specific learning 
on adaptation knowledge acquired by committee members. For this purpose
we conduct experiments on a real tablet formulation problem which is tackled as
a decomposable design task. Evaluation results suggest that adaptation achieves
significant gains compared to a retrieve-only CBR system, but shows that both
learning biases can be beneficial for different decomposed sub-tasks.
References

1.	Bandini, S., Manzoni, S.: CBR adaptation for chemical formulation. In Proc 4th Int Conf on
CBR, pages 634647, 2001. Springer.
2.	Craw, S., Wiratunga, N., Rowe, R.: Case-based design for tablet formulation. In Proc 4th
European Workshop on CBR, pages 358369, 1998. Springer.
3.	Dietterich, T.G.: Ensemble methods in machine learning. In 1st Int Workshop on Multiple
Classifier Systems, pages 115, 2000. Springer.
4.	Domingos, P.: Unifying instance-based and rule-based induction. ML, 24:141168, 1996.
5.	Freund, Y, Schapire, R.: Experiments with a new boosting algorithm. In Proc 13th Int Conf
on Machine Learning, pages 148156, 1996.
6.	Hanney, K., Keane, M.T.: The adaptation knowledge bottleneck: How to ease it by learning
from cases. In Proc of the 2nd Int Conf on CBR, pages 359370, 1997. Springer.
7.	Jarmulak, J., Craw, S., Rowe, R.: Genetic algorithms to optimise CBR retrieval. In Proc 5th
European Workshop on CBR, pages 137149, 2000. Springer.
8.	Jarmulak, J., Craw, S., Rowe, R.: Using case-base data to learn adaptation knowledge for
design. In Proc 17th IJCAI, pages 10111016, 2001. Morgan Kaufmann.
9.	McSherry, D.: An adaptation heuristic for case-based estimation. In Proc 4th European
Workshop on CBR, pages 184-195, 1998. Springer.
10.	McSherry, D.: Minimising dialog length in interactive case-based reasoning. In Proc 17th
IJCAI, pages 993998, 2001. Morgan Kaufmann.
11.	Netten, B.D., Vingerhoeds, R.A.: Incremental adaptation for conceptual design in EADOCS.
In ECAI Workshop on Adaptation in CBR, 1996.
12.	Quinlan, J.R.: Induction of decision trees. Machine Learning, 1:81106, 1986.
13.	Quinlan, J.R.: Bagging, boosting, and C4.5. In Proc 13th National Conf on AI, pages 725730,
1996.
14.	Smyth, B., Cunningham, P.: Dj Vu: A hierarchical case-based reasoning system for software
design. In Proc ECAI92, pages 587589, 1992. Wiley.
15.	Smyth, B., Cunningham, P.: Complexity of adaptation in real-world case-based reasoning
systems. In Proc 6th Irish Conf on Al & Cognitive Science, Ireland, 1993.
16.	Smyth, B., Keane, M.T.: Adaptation-guided retrieval: Questioning the similarity assumption.
Artificial Intelligence, 102:249293, 1998.
17.	Stahl, A., Bergmann, R.: Applying recursive CBR to the customisation of structured products
in an electronic shop. In Proc 5th European Workshop on CBR, pages 297308, 2000.
Springer.
18.	Wettschereck, D., Aha, D.W. (eds.): In Proc ECML-97 Workshop on Case-Based Learn-
ing: Beyond Classification of Feature Vectors, 1997. http://nathan.gmd.de/persons/dietrich.
wettschereck/ecml97-wkshp/ecml97ws2.ps.
19.	Wilke, W., Vollrath, I., Althoff, K.D., Bergmann, R.: A framework for learning adaptation
knowledge based on knowledge light approaches. In 5th German Workshop on CBR, 1997.
