FLOWGEN: Fast and slow graph generation

07/15/2022
by   Aman Madaan, et al.
0

We present FLOWGEN, a graph-generation model inspired by the dual-process theory of mind that generates large graphs incrementally. Depending on the difficulty of completing the graph at the current step, graph generation is routed to either a fast (weaker) or a slow (stronger) model. fast and slow models have identical architectures, but vary in the number of parameters and consequently the strength. Experiments on real-world graphs show that ours can successfully generate graphs similar to those generated by a single large model in a fraction of time.

READ FULL TEXT
research
09/03/2023

An Accurate Graph Generative Model with Tunable Features

A graph is a very common and powerful data structure used for modeling c...
research
06/06/2020

SHADOWCAST: Controlling Network Properties to Explain Graph Generation

We introduce the problem of explaining graph generation, formulated as c...
research
06/05/2019

GRAM: Scalable Generative Models for Graphs with Graph Attention Mechanism

Graphs are ubiquitous real-world data structures, and generative models ...
research
03/13/2022

A Survey on Deep Graph Generation: Methods and Applications

Graphs are ubiquitous in encoding relational information of real-world o...
research
12/10/2021

SLUGGER: Lossless Hierarchical Summarization of Massive Graphs

Given a massive graph, how can we exploit its hierarchical structure for...
research
06/08/2011

Moment based estimation of stochastic Kronecker graph parameters

Stochastic Kronecker graphs supply a parsimonious model for large sparse...
research
02/17/2023

Unsupervised Task Graph Generation from Instructional Video Transcripts

This work explores the problem of generating task graphs of real-world a...

Please sign up or login with your details

Forgot password? Click here to reset