Subjective Attributes in Conversational Recommendation Systems: Challenges and Opportunities

The ubiquity of recommender systems has increased the need for higher-bandwidth, natural and efficient communication with users. This is increasingly filled by recommenders that support language interaction, often conversationally. Given inherent semantic subjectivity present in language, we argue modeling subjective attributes a critical, yet understudied, avenue AI research. We propose novel framework understanding different forms subjectivity, examine various tasks will benefit from systematic treatment attributes, outline number research challenges.