Graph Learning Algorithm

From GM-RKB
Jump to navigation Jump to search

A Graph Learning Algorithm is a graph creation algorithm that is a learning algorithm and can be applied by a graph learning system (to solve a graph learning task).



References

2023

  • https://huggingface.co/blog/intro-graphml
    • QUOTE: The usual process to work on graphs with machine learning is first to generate a meaningful representation for your items of interest (nodes, edges, or full graphs depending on your task), then to use these to train a predictor for your target task. We want (as in other modalities) to constrain the mathematical representations of your objects so that similar objects are mathematically close. However, this similarity is hard to define strictly in graph ML: for example, are two nodes more similar when they have the same labels or the same neighbours?

2021

2018

2017

2017

  • (Egilmez et al., 2017) ⇒ H.E. Egilmez, E. Pavez, A. Ortega. (2017). “Graph Learning from Data under Laplacian and Structural Constraints." In: IEEE Journal of Selected ….
    • QUOTE: “… Since the main focus of the present paper is on formulation of graph learning problems and development of new algorithms, related applications are considered as part of our future work…”
    • NOTE: It emphasizes the need for considering Laplacian and structural constraints in graph learning.