The effective tailoring of decisions to the needs and desires of specific users requires automated mechanisms for preference assessment. We provide a brief overview of recent direct preference elicitation methods: these methods ask users to answer (ideally, a small number of) queries regarding their preferences and use this information to recommend a feasible decision that would be (approximately) optimal given those preferences. We argue for the importance of assessing numerical utilities rather than qualitative preferences, and survey several utility elicitation techniques from artificial intelligence, operations research, and conjoint analysis.
Preference elicitation is a central problem in AI, and has received significant attention in single-agent settings. It is also a key problem in multiagent systems, but has received little attention here so far. In this setting, the agents may have different preferences that often must be aggregated using voting. This leads to interesting issues because what, if any, information should be elicited from an agent depends on what other agents have revealed about their preferences so far. In this paper we study effective elicitation, and its impediments, for the most common voting protocols.
Rational decision making requires full knowledge of the utility function of the person affected by the decisions. However, in many cases, the task of acquiring such knowledge is not feasible due to the size of the outcome space and the complexity of the utility elicitation process. Given that the amount of utility information we can acquire is limited, we need to make decisions with partial utility information and should carefully select which utility elicitation questions we ask. In this paper, we propose a new approach for this problem that utilizes a prior probability distribution over the person's utility function, perhaps learned from a population of similar people. The relevance of a utility elicitation question for the current decision problem can then be measured using its value of information. We propose an algorithm that interleaves the analysis of the decision problem and utility elicitation to allow these two tasks to inform each other. At every step, it asks the utility elicitation question giving us the highest value of information and computes the best strategy based on the information acquired so far, stopping when the expected utility loss resulting from our recommendation falls below a pre-specified threshold. We show how the various steps of this algorithm can be implemented efficiently.
Bayesian approaches to preference elicitation (PE) are particularly attractive due to their ability to explicitly model uncertainty in users' latent utility functions. However, previous approaches to Bayesian PE have ignored the important problem of generalizing from previous users to an unseen user in order to reduce the elicitation burden on new users. In this paper, we address this deficiency by introducing a Gaussian Process (GP) prior over users' latent utility functions on the joint space of user and item features. We learn the hyper-parameters of this GP on a set of preferences of previous users and use it to aid in the elicitation process for a new user. This approach provides a flexible model of a multi-user utility function, facilitates an efficient value of information (VOI) heuristic query selection strategy, and provides a principled way to incorporate the elicitations of multiple users back into the model.
Most frameworks for utility elicitation assume a predefined set of features over which user preferences are expressed. We consider utility elicitation in the presence of subjective or user-defined features, whose definitions are not known in advance. We treat the problem of learning a user's feature definition as one of concept learning, but whose goal is to learn only enough about the concept definition to enable a good decision to be made. This is complicated by the fact that user utility is unknown. We describe computational procedures for identifying optimal alternatives w.r.t minimax regret in the presence of both utility and concept uncertainty; and develop several heuristic query strategies that focus simultaneously on reduction of relevant concept and utility uncertainty.