Description
To join by Zoom : https://uw-edu-pl.zoom.us/j/91603361707
Meeting ID: 916 0336 1707
Speaker: Rafał Masełek
Title: Machine Learning Electroweakino Production
Abstract:
The nature of dark matter remains one of the universe's most profound mysteries. If dark matter interacts feebly with Standard Model particles, it could be produced and detected at the Large Hadron Collider. The monojet channel is particularly intriguing, where one or a few high-energy jets recoil against missing transverse momentum without isolated leptons. Traditional analyses often struggle with the low-level information crucial for this channel. In this talk, I will discuss my recent work, "Machine Learning Electroweakino Production" (https://arxiv.org/abs/2411.00093) , which introduces a novel analysis for the monojet channel using Graph Neural Networks applied to searches for electroweakinos. Our approach combines both reconstructed objects and low-level particle-flow information, enhancing signal-to-background discrimination. I will present the accuracy of our machine learning approach, its sensitivity to new physics, and discuss the interpretability of our network, shedding light on the physical features it identifies for effective discrimination.
The nature of dark matter remains one of the universe's most profound mysteries. If dark matter interacts feebly with Standard Model particles, it could be produced and detected at the Large Hadron Collider. The monojet channel is particularly intriguing, where one or a few high-energy jets recoil against missing transverse momentum without isolated leptons. Traditional analyses often struggle with the low-level information crucial for this channel. In this talk, I will discuss my recent work, "Machine Learning Electroweakino Production" (https://arxiv.org/abs/2411.00093) , which introduces a novel analysis for the monojet channel using Graph Neural Networks applied to searches for electroweakinos. Our approach combines both reconstructed objects and low-level particle-flow information, enhancing signal-to-background discrimination. I will present the accuracy of our machine learning approach, its sensitivity to new physics, and discuss the interpretability of our network, shedding light on the physical features it identifies for effective discrimination.
Organised by
Rafal Maselek, Sabine Kraml
Contact to R. Maselek