A recommender system (RS) can infer constraints on the user utility function by observing the queries selected by a user among those it has suggested. Reasoning on these constraints it can avoid suggesting queries that retrieve products with an inferior utility, i.e., dominated queries. In this paper we propose a new efficient technique for the computation of dominated queries. It relies on the system's assumption that the number of possible profiles (utility functions) of the users it may interact with is finite. Under this assumption query suggestions can be efficiently computed and their number can be kept small. Moreover, we show that even if the system is not contemplating all the possible user profiles its performance is very close to the optimal one.

Query revisions in a conversational system can be efficiently computed by assuming that the profiles of the potential users are in a predefined, a priori known and finite set. However, without any additional knowledge of the actual profiles distribution, the system may miss the true profiles of the users, hence deteriorating the system performance. We propose a method for identifying a tailored set of profiles that is acquired by analysing the implicitly shown preferences of the users that interacted with the system. We show that with the proposed method the system can efficiently identify good query revisions.

Describes a unique experiment in safety management which has led to a significant improvement in the safety performance of work groups at the smelter of INCO Ltd in Sudbury, Canada.

No abstract