A master thesis from Aalborg University
Performance Aware Layer Combination for Graph Convolutional Network
Term
4. term
Education
Publication year
2021
Submitted on
2021-06-10
Pages
32 pages
Abstract
Graph Convolutional Network (GCN) is a state-of-the-art method used for recommendation. Throughout this paper we study the effects of modifying the methods used for layer combination in GCN. An ablation study for GCF is conducted to understand why it outperformed LightGCN on their datasets. We focus on LightGCN which is a simplified implementation of the Neural Graph Collaborative Filtering (NGCF). LightGCN outperformed NGCF and other state-of-the-art methods. We propose two new extensions for LightGCN called Aggressive Layer Combination (ALC) and Balanced Layer Combination (BLC) instead of LightGCN's version of weighted summation for layer combination. This showed better results on most datasets compared to both GCF and LightGCN. We also show that in certain cases only utilizing the embedding from a single layer showed to outperform ALC, BLC, GCF and LightGCN.
Keywords
recommendation ; graph ; convolutional ; network ; GCN ; lightgcn ; ngcf ; gcf ; bitgcf ; collaborative filtering
Documents
Colophon: This page is part of the AAU Student Projects portal, which is run by Aalborg University. Here, you can find and download publicly available bachelor's theses and master's projects from across the university dating from 2008 onwards. Student projects from before 2008 are available in printed form at Aalborg University Library.
If you have any questions about AAU Student Projects or the research registration, dissemination and analysis at Aalborg University, please feel free to contact the VBN team. You can also find more information in the AAU Student Projects FAQs.