An Experimental Assessment of the Stability of Graph Contrastive Learning

The Deep Neural Network (DNN) has become a widely popular machine learning architecture thanks to its ability to learn complex behaviors from data. Standard learning strategies for DNNs however rely on the availability of large, labeled datasets. Self-Supervised Learning (SSL) is a style of learning that allows models to also use unlabeled data...

Subgraph Matching via Fused Gromov-Wasserstein Distance

Subgraph matching is a fundamental problem in various fields such as machine learning, computer vision, image processing, and bioinformatics, where detecting specific substructures within an object is often crucial. In these domains, not only structure plays an essential role, but also the feature information on nodes should be incorporated,...

Architectural Pipeline - Pipeline Architecture: An experiment into the role of topological graphs in the early stages of architectural design in the era of machine learning

The realm of architectural design and conceptualization, despite witnessing advancements in design complexity facilitated by technological tools and fabrication techniques, appears to have experienced limited transformative change over time in the way we begin our design process. But is there an alternative way to look at architecture ? Perhaps...

Quality assessment of a graph-based approach for a multi-methods determination of geotechnical model parameters

The determination of geotechnical parameters from in-situ tests heavily relies on the use of empirical correlations relating field measurements to soil properties. A large number of these correlations can be found across the literature, each of them yielding a different outcome for the parameter value. This results in a high variability of the...

Analysis Of Gas Transport Networks: Strategies For Simplifying And Solving Problems Related To Gas Transport Networks

In this research, known theory and methods for electric networks are translated and applied to gas transport networks.<br/><br/>The networks are represented by directed graphs with their corresponding incidence matrices. The theory about these is extensively discussed in addition to their relevance to Kirchoff's Circuit Laws.<br/><br/>Several...

Automatic Generation of Chemical Reactor Networks for Combustion Simulations

Gas turbines are the power house of most modern day aircrafts. Combustion of fossil fuels provides the energy required, but this produces harmful emissions. With ever increasing strictness on emissions regulation, it is becoming even more necessary to design gas turbine combustion chambers with the intent of low pollutant output. Gas turbine...

Blind Graph Topology Change Detection: A Graph Signal Processing approach

Graphs are used to model irregular data structures and serve as models to represent/capture the interrelationships between data. The data in graphs are also referred as graph signals. Graph signal processing (GSP) can then be applied which basically extends classical signal processing to solve problems. Anomaly detection is an example of such a...

Graphalytics Global Competition: A Competition Platform to Compare Different Graph Processing Platforms

Graphs are becoming more popular day by day. This has lead to the development of different graph-processing platforms, such as Hadoop, Apache Giraph and GraphLab. With this wide variety of graph processing platforms, which one should we choose for a specific use case? Which platform has the best performance? To solve this issue, we can benchmark...