Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Effective Two-Stage Knowledge Transfer for Multi-Entity Cross-Domain Recommendation

Jianyu Guan, Zongming Yin, Tianyi Zhang, Leihui Chen, Yin Zhang, Fei Huang, Jufeng Chen, Shuguang Han

TL;DR

A pre-training & fine-tuning based Multi-entity Knowledge Transfer framework called MKT, which demonstrates the superiority of MKT over multiple State-Of-The-Art methods for multi-entity cross-domain recommendation.

Abstract

In recent years, the recommendation content on e-commerce platforms has become increasingly rich -- a single user feed may contain multiple entities, such as selling products, short videos, and content posts. To deal with the multi-entity recommendation problem, an intuitive solution is to adopt the shared-network-based architecture for joint training. The idea is to transfer the extracted knowledge from one type of entity (source entity) to another (target entity). However, different from the conventional same-entity cross-domain recommendation, multi-entity knowledge transfer encounters several important issues: (1) data distributions of the source entity and target entity are naturally different, making the shared-network-based joint training susceptible to the negative transfer issue, (2) more importantly, the corresponding feature schema of each entity is not exactly aligned (e.g., price is an essential feature for selling product while missing for content posts), making the existing methods no longer appropriate. Recent researchers have also experimented with the pre-training and fine-tuning paradigm. Again, they only consider the scenarios with the same entity type and feature systems, which is inappropriate in our case. To this end, we design a pre-training & fine-tuning based Multi-entity Knowledge Transfer framework called MKT. MKT utilizes a multi-entity pre-training module to extract transferable knowledge across different entities. In particular, a feature alignment module is first applied to scale and align different feature schemas. Afterward, a couple of knowledge extractors are employed to extract the common and entity-specific knowledge. In the end, the extracted common knowledge is adopted for target entity model training. Through extensive offline and online experiments, we demonstrated the superiority of MKT over multiple State-Of-The-Art methods.

Effective Two-Stage Knowledge Transfer for Multi-Entity Cross-Domain Recommendation

TL;DR

A pre-training & fine-tuning based Multi-entity Knowledge Transfer framework called MKT, which demonstrates the superiority of MKT over multiple State-Of-The-Art methods for multi-entity cross-domain recommendation.

Abstract

In recent years, the recommendation content on e-commerce platforms has become increasingly rich -- a single user feed may contain multiple entities, such as selling products, short videos, and content posts. To deal with the multi-entity recommendation problem, an intuitive solution is to adopt the shared-network-based architecture for joint training. The idea is to transfer the extracted knowledge from one type of entity (source entity) to another (target entity). However, different from the conventional same-entity cross-domain recommendation, multi-entity knowledge transfer encounters several important issues: (1) data distributions of the source entity and target entity are naturally different, making the shared-network-based joint training susceptible to the negative transfer issue, (2) more importantly, the corresponding feature schema of each entity is not exactly aligned (e.g., price is an essential feature for selling product while missing for content posts), making the existing methods no longer appropriate. Recent researchers have also experimented with the pre-training and fine-tuning paradigm. Again, they only consider the scenarios with the same entity type and feature systems, which is inappropriate in our case. To this end, we design a pre-training & fine-tuning based Multi-entity Knowledge Transfer framework called MKT. MKT utilizes a multi-entity pre-training module to extract transferable knowledge across different entities. In particular, a feature alignment module is first applied to scale and align different feature schemas. Afterward, a couple of knowledge extractors are employed to extract the common and entity-specific knowledge. In the end, the extracted common knowledge is adopted for target entity model training. Through extensive offline and online experiments, we demonstrated the superiority of MKT over multiple State-Of-The-Art methods.
Paper Structure (23 sections, 14 equations, 3 figures, 4 tables)

This paper contains 23 sections, 14 equations, 3 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Methodology
  4. Model Overview
  5. Multi-Entity Model Pre-training
  6. Feature Embedding
  7. Heterogeneous Feature Alignment(HFA)
  8. Common Knowledge Extractor (CKE)
  9. Polarized Distribution Loss (PDL)
  10. Target-Entity Model Fine-tuning
  11. Online Deployment
  12. Experiment
  13. Experiment Setup
  14. Datasets
  15. Baselines
  16. ...and 8 more sections

Figures (3)

  • Figure 1: Homepage recommendation of Xianyu APP, which consists of a mixture of selling products and content postings.
  • Figure 2: An overview of the model architecture for MKT, which consists of a multi-entity model (left) that is trained on the mixed-domain data, and a target entity model (right) for the target entity recommendation. Modules in the gray color denote the online serving part.
  • Figure 3: An illustration of the HFA component.