Retrieval Issues in Real-World CBR Applications How far can we go with discrimination-nets?

Abstract

We propose that analogical reasoning and case based reasoning (CBR) tasks can be usefully characterised as a continuum reflecting the remoteness of the remindings involved. Remindings in CBR are generally between semantically close cases while analogical reasoning depends on more abstract remindings. Rather than there being a strict demarcation between CBR and analogical reasoning on these grounds (with analogical reasoning concerned with inter domain remindings and CBR dealing with remindings within one domain) there is a continuum of cognitive tasks that draw on past experience during reasoning. Simpler tasks like diagnosis and classification are located near the CBR end while more complex tasks like creative design are located towards the analogical reasoning end. The question is how far towards the abstract end of the continuum can the index-based retrieval techniques that are effective in CBR be used (eg. discrimination networks). We are considering episode retrieval as a two stage process; the first stage being the initial filtering of the case base, and the second stage selecting the best case from this candidate set. We focus on the base filtering stage and conclude that discrimination networks are adequate for comparatively complex cognitive tasks such as routine design. However, we argue that CBR systems for non-routine design should provide interactive case retrieval and act as CBR assistants.