English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

The problem of labeled graph generation is gaining attention in the DeepLearning community. The task is challenging due to the sparse and discretenature of graph spaces. Several approaches have been proposed in theliterature, most of which require to transform the graphs into sequences thatencode their structure and labels and to learn the distribution of suchsequences through an auto-regressive generative model. Among this family ofapproaches, we focus on the GraphGen model. The preprocessing phase of GraphGentransforms graphs into unique edge sequences called Depth-First Search (DFS)codes, such that two isomorphic graphs are assigned the same DFS code. Eachelement of a DFS code is associated with a graph edge: specifically, it is aquintuple comprising one node identifier for each of the two endpoints, theirnode labels, and the edge label. GraphGen learns to generate such sequencesauto-regressively and models the probability of each component of the quintupleindependently. While effective, the independence assumption made by the modelis too loose to capture the complex label dependencies of real-world graphsprecisely. By introducing a novel graph preprocessing approach, we are able toprocess the labeling information of both nodes and edges jointly. Thecorresponding model, which we term GraphGen-Redux, improves upon the generativeperformances of GraphGen in a wide range of datasets of chemical and socialgraphs. In addition, it uses approximately 78% fewer parameters than thevanilla variant and requires 50% fewer epochs of training on average.

GraphGen-Redux: a Fast and Lightweight Recurrent Model for labeled Graph Generation