Seminar at NAVER LABS Europe: Recent advances in document network embedding

Abstract : Many sources of our informational landscape can be formalized as a network of documents. For a long time the textual content of documents and the structure that shows how documents relate to each other have been considered separately. Recently document network embedding has been proposed to learn representations that take both content and structure into account. This space can then be used for downstreams tasks, such as classification or link prediction. In this talk I will give an overview of recent methods that aim at building such embedding space. In particular, I will focus on several models that were recently proposed in the ERIC Lab [1,2,3,4].

[1] R. Brochier, A. Guille and J. Velcin (2019). Global Vectors for Node Representation. Proceedings of the International World Wide Web Conference (WWW), May 13–17, 2019, San Francisco, CA, USA.

[2] R. Brochier, Guille and J. Velcin (2019). Link Prediction with Mutual Attention for Text-Attributed Networks. Proceedings of the International Workshop on Deep Learning for Graphs and Structured Data Embedding, colocated with WWW (Companion Volume), May 13–17, 2019, San Francisco, CA, USA.

[3] A. Gourru, J. Velcin, J. Jacques and A. Guille (2020). Document Network Projection in Pretrained Word Embedding Space. Paper accepted at ECIR 2020, Lisbon, Portugal.

[4] R. Brochier, A. Guille and J. Velcin (2020). Inductive Document Network Embedding with Topic-Word Attention. Paper accepted at ECIR 2020, Lisbon, Portugal.