Ensemble Node Embeddings using Tensor Decomposition: A Case-Study on DeepWalk

Authors

Jia Chen, The University of Texas Rio Grande Valley
Evangelos E. Papalexakis

Document Type

Article

Publication Date

2020

Abstract

Node embeddings have been attracting increasing attention during the past years. In this context, we propose a new ensemble node embedding approach, called TENSEMBLE2VEC, by first generating multiple embeddings using the existing techniques and taking them as multiview data input of the state-of-art tensor decomposition model namely PARAFAC2 to learn the shared lower-dimensional representations of the nodes. Contrary to other embedding methods, our TENSEMBLE2VEC takes advantage of the complementary information from different methods or the same method with different hyper-parameters, which bypasses the challenge of choosing models. Extensive tests using real-world data validates the efficiency of the proposed method.

Recommended Citation

Chen, J., & Papalexakis, E. E. (2020). Ensemble Node Embeddings using Tensor Decomposition: A Case-Study on DeepWalk. arXiv preprint arXiv:2008.07672.