Next: Programmed and learned Up: Belew: Research Statement Previous: Introduction

Adaptive information retrieval

In my thesis I first began to investigate the application of (what is now called) ``connectionist'' learning techniques to the IR problem with a system called AIR [16]. This has proven to be a very appropriate marriage of technology with application, as evidenced by subsequent progress by me, my students, and now quite a number of others. In particular I continue to enjoy a rich collaboration with Dan Rose since advising his thesis [42]; a slightly edited version of Dan's thesis has been republished as a book by Lawrence Erlbaum [43].

My own research has focused on three aspects of AIR, described below from the perspectives of the AI, IR and (loosely) social science perspectives, resp.