Title: Graph Neural Network: Its applications to constrain BSM models and EFTs
Abstract:
Graph Neural Networks have emerged as a powerful tool for operating on graph-structured data, facilitating the exploration of non-Euclidean physics data. In this talk, I will first discuss an autoencoder-based strategy to facilitate anomaly detection for boosted jets, employing Graph Neural Networks (GNNs). To overcome the known limitations of GNN autoencoders, we design a symmetric decoder capable of simultaneously reconstructing edge features and node features.
Next, I will discuss the application of GNNs in a supervised scenario where we explore its potential to improve high-dimensional effective field theory parameter fits to collider data beyond traditional rectangular cut-based differential distribution analyses. As a specific case, we focus on an SMEFT analysis of pp → top pair production, including top decays, where the linear effective field deformation is parameterized by thirteen independent Wilson coefficients. The application of GNNs allows us to condense the multidimensional phase space information available for the discrimination of BSM effects from the SM expectation by considering all available final state correlations directly.
Zoom Meeting ID: 948 7183 3595