Spatial-Temporal Knowledge Distillation for Takeaway Recommendation

TL;DR

This paper introduces a novel spatial-temporal knowledge distillation model for takeaway recommendation that effectively captures dynamic user preferences and integrates complex geospatial data with low computational costs.

Contribution

It proposes a two-stage training framework combining a spatial-temporal knowledge graph encoder and a Transformer, enabling efficient fusion of graph and sequence data for improved recommendations.

Findings

Outperforms state-of-the-art baselines on three datasets

Effectively models dynamic user preferences and geospatial information

Reduces computational costs compared to existing methods

Abstract

The takeaway recommendation system aims to recommend users' future takeaway purchases based on their historical purchase behaviors, thereby improving user satisfaction and boosting merchant sales. Existing methods focus on incorporating auxiliary information or leveraging knowledge graphs to alleviate the sparsity issue of user purchase sequences. However, two main challenges limit the performance of these approaches: (1) capturing dynamic user preferences on complex geospatial information and (2) efficiently integrating spatial-temporal knowledge from both graphs and sequence data with low computational costs. In this paper, we propose a novel spatial-temporal knowledge distillation model for takeaway recommendation (STKDRec) based on the two-stage training process. Specifically, during the first pre-training stage, a spatial-temporal knowledge graph (STKG) encoder is trained to extract high-order spatial-temporal dependencies and collaborative associations from the STKG. During the second spatial-temporal knowledge distillation (STKD) stage, a spatial-temporal Transformer (ST-Transformer) is employed to comprehensively model dynamic user preferences on various types of fine-grained geospatial information from a sequential perspective. Furthermore, the STKD strategy is introduced to transfer graph-based spatial-temporal knowledge to the ST-Transformer, facilitating the adaptive fusion of rich knowledge derived from both the STKG and sequence data while reducing computational overhead. Extensive experiments on three real-world datasets show that STKDRec significantly outperforms the state-of-the-art baselines.