A Socially-Aware Conversational Recommender System for Personalized Recipe Recommendations

One potential solution to help people change their eating behavior is to develop conversational systems able to recommend healthy recipes. Beyond the intrinsic quality of the recommendations themselves, various factors might also influence users’ perception of a recommendation. Two of these factors are the conversational skills of the system and users’ interaction modality. In this paper, we present Cora, a conversational system that recommends recipes aligned with its users’ eating habits and current preferences. Users can interact with Cora in two different ways. They can select predefined answers by clicking on buttons to talk to Cora or write text in natural language. On the other hand, Cora can engage users through a social dialogue, or go straight to the point. We conduct an experiment to evaluate the impact of Cora’s conversational skills and users’ interaction mode on users’ perception and intention to cook the recommended recipes. Our results show that a conversational recommendation system that engages its users through a rapport-building dialogue improves users’ perception of the interaction as well as their perception of the system.

Introduction. Healthy eating implies complex decision making processes [6], including being aware of healthy options and choosing among them [24]. One solution to overcome this issue and help people to make healthier choices is to develop health-aware food recommender systems [31]. While significant effort has been put recently into optimizing the food selection algorithms [30], many other factors can also influence users’ overall experience when interacting with a recommender system [14]. Indeed, the way the recommendation is presented [18], the system’s response time [33], or even the length of the system’s utterances [20] can have an influence on users’ perception of the system. One trend to improve users’ experience is to make the interaction more natural by designing the recommendation process as a conversation [23]. Besides helping users to achieve task-oriented goals, conversations can also fulfill interpersonal functions, such as building rapport [29].

Discussion / Conclusion. The good ratings obtained across each condition combined with the high acceptance rate show that participants have been generally satisfied with Cora and its recommendations. Regardless of the quality of the recommendations, our results also show that endowing recommender systems with rapport-building abilities has a positive influence on users’ perception. That is corroborated by the fact that the rapport-building version of Cora systematically obtained better scores than its task-oriented counterpart, and lower standard deviations. In other words, participants preferred the rapport-building version of Cora, and their ratings of this version were more consistent. Furthermore, not only are participants significantly more willing to use a system able to engage them in a rapport-building conversation, but they also perceive that a rapport-building system delivers significantly more details about the recommendations, although it simply gives its "own" personal opinion.