Date of Award
Santa Clara : Santa Clara University, 2022.
Doctor of Philosophy (PhD)
Computer Science and Engineering
Dialogue systems are becoming an increasingly common part of many users' daily routines. Natural language serves as a convenient interface to express our preferences with the underlying systems. In this work, we implement a full-fledged Conversational Recommendation System, mainly focusing on learning user preferences through online conversations. Compared to the traditional collaborative filtering setting where feedback is provided quantitatively, conversational users may only indicate their preferences at a high level with inexact item mentions in the form of natural language chit-chat. This makes it harder for the system to correctly interpret user intent and in turn provide useful recommendations to the user. To tackle the ambiguities in natural language conversations, we propose Personalized Memory Transfer (PMT) which learns a personalized model in an online manner by leveraging a key-value memory structure to distill user feedback directly from conversations. This memory structure enables the integration of prior knowledge to transfer existing item representations/preferences and natural language representations. We also implement a retrieval based response generation module, where the system in addition to recommending items to the user, also responds to the user, either to elicit more information regarding the user intent or just for a casual chit-chat. The experiments were conducted on two public datasets and the results demonstrate the effectiveness of the proposed approach.
Godavarthy, Naga Archana, "Personalized Memory Transfer for Conversational Recommendation Systems" (2022). Engineering Ph.D. Theses. 39.