What are Dynamic Graphs and Why They are Interesting?

Original Source Here

Use of Dynamic Graphs

Photo by Emily Morter on Unsplash

As I mentioned before, the dynamic graphs are useful in capturing the complex behaviour of real-world entities (like human beings in a social network, or a gene coding protein in a PPI network). We also add the time dimension to the dataset which explains the evolution of a process from start to finish.

In order to extract valuable information from dynamic graph data, we can employ machine learning techniques. Firstly, we need to learn numerical representations of the entities present in the dynamic graph and then we can use methods such as Graph Neural Networks to model several use cases like node classification, link prediction etc.

One very common task on such graphs is community detection, which in the context of PPI is known as complex detection. Basically, we use unsupervised machine learning techniques such as clustering to identify important groups at each timepoint. For the PPI network, this is interesting because the identification of important protein clusters/ complexes would result in better drug discovery for diseases.

We also have a graph alignment task where we aim to identify proteins or complexes that show variations among different types of samples (like healthy or infected cells). This is useful for biologists in finding the proteins on which they should focus when looking for treatments.

In the case of social networks, we can use dynamic graph analysis to identify the spammers in the network (spam detection). The core idea of dynamic graphs is the same but different domains have their own use cases and we can use machine learning models to cater for their needs.

Using Link Prediction, we can propose friend recommendations to users in a social network by analysing the growth of the network over time. For PPIs, we can provide a list of protein recommendations, meaning we want to propose a list of proteins that have the potential to form new interactions.


Trending AI/ML Article Identified & Digested via Granola by Ramsey Elbasheer; a Machine-Driven RSS Bot

%d bloggers like this: