Text-Conditioned Graph Generation Using Discrete Graph Variational Autoencoders

Michael Longland, David Liebowitz, Kristen Moore, Salil S. Kanhere

Abstract: Inspired by recent progress in text-conditioned image generation, we propose a model for the novel problem of text-conditioned graph generation. In this paper we introduce the Vector Quantized Text To Graph generator (VQ-T2G), a discrete graph variational autoencoder and autoregressive transformer for generating general graphs conditioned on text. We curate two multimodal datasets of graphs paired with text, a real-world dataset of 8000 subgraphs from the Wikipedia link network and a dataset of over 5000 synthetic graphs. Experimental results on these datasets demonstrate that VQ-T2G synthesises novel graphs with structure aligned with the text conditioning. Additional experiments in the unconditioned graph generation setting show VQ-T2G is competitive with existing unconditioned graph generation methods across various graph metrics. Code will be released on github following paper acceptance.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Submission Guidelines: Yes

Please Choose The Closest Area That Your Submission Falls Into: Generative models