TY - GEN
T1 - Developing a Conversational Recommendation Systemfor Navigating Limited Options
AU - Bursztyn, Victor S.
AU - Healey, Jennifer
AU - Koh, Eunyee
AU - Lipka, Nedim
AU - Birnbaum, Larry
N1 - Publisher Copyright:
© 2021 ACM.
PY - 2021/5/8
Y1 - 2021/5/8
N2 - We have developed a conversational recommendation system designed to help users navigate through a set of limited options to find the best choice. Unlike many internet scale systems that use a singular set of search terms and return a ranked list of options from amongst thousands, our system uses multi-turn user dialog to deeply understand the user's preferences. The system responds in context to the user's specific and immediate feedback to make sequential recommendations. We envision our system would be highly useful in situations with intrinsic constraints, such as finding the right restaurant within walking distance or the right retail item within a limited inventory. Our research prototype instantiates the former use case, leveraging real data from Google Places, Yelp, and Zomato. We evaluated our system against a similar system that did not incorporate user feedback in a 16 person remote study, generating 64 scenario-based search journeys. When our recommendation system was successfully triggered, we saw both an increase in efficiency and a higher confidence rating with respect to final user choice. We also found that users preferred our system (75%) compared with the baseline.
AB - We have developed a conversational recommendation system designed to help users navigate through a set of limited options to find the best choice. Unlike many internet scale systems that use a singular set of search terms and return a ranked list of options from amongst thousands, our system uses multi-turn user dialog to deeply understand the user's preferences. The system responds in context to the user's specific and immediate feedback to make sequential recommendations. We envision our system would be highly useful in situations with intrinsic constraints, such as finding the right restaurant within walking distance or the right retail item within a limited inventory. Our research prototype instantiates the former use case, leveraging real data from Google Places, Yelp, and Zomato. We evaluated our system against a similar system that did not incorporate user feedback in a 16 person remote study, generating 64 scenario-based search journeys. When our recommendation system was successfully triggered, we saw both an increase in efficiency and a higher confidence rating with respect to final user choice. We also found that users preferred our system (75%) compared with the baseline.
KW - agreement
KW - conversational
KW - interactive
KW - natural language processing
KW - recommendation system
UR - http://www.scopus.com/inward/record.url?scp=85105820468&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85105820468&partnerID=8YFLogxK
U2 - 10.1145/3411763.3451596
DO - 10.1145/3411763.3451596
M3 - Conference contribution
AN - SCOPUS:85105820468
T3 - Conference on Human Factors in Computing Systems - Proceedings
BT - Extended Abstracts of the 2021 CHI Conference on Human Factors in Computing Systems, CHI EA 2021
PB - Association for Computing Machinery
T2 - 2021 CHI Conference on Human Factors in Computing Systems: Making Waves, Combining Strengths, CHI EA 2021
Y2 - 8 May 2021 through 13 May 2021
ER -