User Memory Reasoning for Conversational Recommendation
Abstract
We study a conversational recommendation model which dynamically manages users' past (offline) preferences and current (online) requests through a structured and cumulative user memory knowledge graph, to allow for natural interactions and accurate recommendations. For this study, we create a new Memory Graph (MG) <--> Conversational Recommendation parallel corpus called MGConvRex with 7K+ human-to-human role-playing dialogs, grounded on a large-scale user memory bootstrapped from real-world user scenarios. MGConvRex captures human-level reasoning over user memory and has disjoint training/testing sets of users for zero-shot (cold-start) reasoning for recommendation. We propose a simple yet expandable formulation for constructing and updating the MG, and a reasoning model that predicts optimal dialog policies and recommendation items in unconstrained graph space. The prediction of our proposed model inherits the graph structure, providing a natural way to explain the model's recommendation. Experiments are conducted for both offline metrics and online simulation, showing competitive results.
- Publication:
-
arXiv e-prints
- Pub Date:
- May 2020
- DOI:
- 10.48550/arXiv.2006.00184
- arXiv:
- arXiv:2006.00184
- Bibcode:
- 2020arXiv200600184X
- Keywords:
-
- Computer Science - Computation and Language