Applications of Machine Learning in Probing Top-Quark Flavor-Changing Interactions and Jet Charge Discrimination

• Meisam Ghasemi (School of Particles and Accelerators - IPM)

Machine learning is becoming an essential tool in high-energy physics for improving event classification and uncovering rare phenomena. In this talk, I present two applications: a search for top-quark flavor-changing neutral current interactions within the Effective Field Theory framework, and a jet charge discrimination study using modern deep learning methods. In the first, deep neural networks enhance sensitivity to non-SM top interactions, leading to stronger exclusion limits on the new physics scale. In the second, graph neural networks outperform traditional models by exploiting correlations among charged particles to accurately infer the initiating quark’s electric charge.