Conversational Group Recommender Systems

T. Nguyen
DOI: https://doi.org/10.1145/3079628.3079704
2017-07-09
Abstract:Recommending to a group of users is multifaceted as people naturally adapt to other members, and it may turn out that what they choose in a group does not fully match individual interests. Besides, it has been shown that the recommendation needs of groups go beyond the aggregation of individual preferences. In practice, it is much more difficult to predict group choices because users take into account the others' reactions and different users react to the group in different ways. Thus, in this research, we aim at exploiting an interactive and conversational approach to facilitate the group decision-making process where the complex trade-off between the satisfaction of an individual and the group as a whole typically occurs and needs to be resolved. To attain this goal, we investigate approaches that can access a group situation and autonomously learn an adaptive interaction in a specific condition of the group.
What problem does this paper attempt to address?