Abstract:
Over the past few decades, significant advances have been made in understanding the fundamental properties of neutrinos, leading to groundbreaking measurements and discoveries. However, challenges persist, including to determine the CP-violation phase $\delta_{\mathrm{CP}}$, which remains experimentally elusive. The leading causes of these challenges include the complexity of systematic uncertainties from various sources, which intricately influence the observed physics in neutrino experiments. In a conventional physics analysis pipeline, the understanding of detector responses often relies on empirically derived assumptions, leading to separate calibrations targeting various effects. The time-consuming nature of this approach can limit the timely analysis upgrades. Moreover, it lacks the adaptability to accommodate discrepancies arising from asymptotic inputs and factorized physics processes.
In recent years, AI and machine learning (ML) techniques have achieved remarkable success across a variety of applications and research domains, making them strong candidates for addressing the challenges of event reconstruction in modern neutrino experiments. Early applications of AI/ML in neutrino physics demonstrate significant potential for tasks such as feature recognition, fast surrogates, and differentiable simulation. Differentiable physics emulator can enhance the estimations of systematic uncertainties and advances physics inference across the aforementioned challenges. The differentiability helps realize the optimizability of model with data, through which users can infer convoluted detector effects via a single differentiable model, informed by robust physics knowledge inputs.
Biography:
B.S. in Astrophysics at UCLA (2017) ; Ph.D. of Physics at University of Tokyo (2022) ; JSPS PostDoc Fellow at IPMU, Univ. of Tokyo (2022-present). Working on Super-Kaimokande, T2K, and Hyper-Kamiokande since 2017 and now as T2K-SK analysis group convener. Interested in experimental neutrino physics, multi-messenger astrophysics, and AI/ML application in physics.
Host: Prof. Junting Huang
Alternative online link: https://meeting.tencent.com/dm/4k5jPUBMPXhZ (id: 900816701 passcode: 123456）