Learning What to Want: Context-Sensitive Preference Learning

Abstract

We have developed a method for learning relative preferences from histories of choices made, without requiring an intermediate utility computation. Our method infers preferences that are rational in a psychological sense, where agent choices result from Bayesian inference of what to do from observable inputs. We further characterize conditions on choice histories wherein it is appropriate for modelers to describe relative preferences using ordinal utilities, and illustrate the importance of the influence of choice history by explaining all major categories of context effects using them. Our proposal clarifies the relationship between economic and psychological definitions of rationality and rationalizes several behaviors heretofore judged irrational by behavioral economists.

Toward Better Interactions in Recommender Systems

“…From the literature, dynamics can be achieved with two approaches: modelbased and algorithm-based. Model-based approaches include context-aware recommenders [29,1] and systems explicitly modeling user preference change [13,3]. For example, in their work on Context-Aware Recommender Systems (CARS), Adomavicius et al [1] examined how context can be defined and used in order to create more intelligent recommendations, such as using pre-filtering and post-filtering strategies with respect to contextual factors.…”

Context-Sensitive Computational Mechanisms of Decision Making

“…In multialternative choice, inclusion of other options can alter choices among a fixed set of options. 35 , 36 The attraction effect is the enhancement of the preference for one of the options by introducing a similar but inferior decoy option. The similarity effect increases the probability of selecting the dissimilar option, and the compromise effect increases the probability of selecting the third option that is intermediate to the 2 original options.…”

Inducing preference reversals by manipulating revealed preferences

“…The earliest conceptualization of a preference reversal can be traced to Luce & Raiffa's fastidious diner, who initially prefers salmon to steak off of a restaurant menu, but changes his mind and orders steak instead when the waiter tells him that the day's special is frog legs (Luce & Raiffa, 1957). While whimsical, this story points directly to a core objection to the tenability of optionspecific representations of value in the mind (Srivastava & Schrater, 2015). If the diner prefers salmon to steak to begin with, why does the introduction of an additional item shift their preference to steak?…”

“…In other words, the diner infers something about the generative process underlying the options from the set of options, and then uses his understanding of the generative process to construct his preference. Preference reversals and more generally, all such context effects, are deeply interesting because they uncover the existence of such sophisticated inferences underpinning the simple act of choosing between items (Srivastava & Schrater, 2015).…”