Date/Time: Thu Apr 03 2025 at 14:00<br/><br/>Location: Auditorium<br/><br/>Speaker: Luise Poley (TRIUMF)<br/><br/>Title: Cold Noise - a cautionary tale about large projects and small details<br/><br/>Abstract: In 2022, after more than ten years of R&D, the production of 20,000 detector modules for the ATLAS strip tracker was about to start. Instead, tests of recently assembled modules showed excess noise peaks of up to 100% on a large fraction of modules operated at the nominal detector temperature of -35C. Despite multiple stages of prototyping and expert reviews, the problem had not been caught by previous quality control or design verification. Instead of starting production, an investigation was launched to identify the underlying mechanism. It involved fifteen institutes and spanned more than twelve months as every conceivable parameter of the involved components and their interfaces was being investigated. In an unexpected series of events, Cold Noise defeated any and all attempts to predict its behaviour or mitigate its impact, until the effect was traced down to an unexpected interplay between several seemingly unimportant details which were later described as "impossible to produce or design intentionally". Besides its technical aspects, Cold Noise presents an interesting case of how a series of decisions and prioritisations made with the best intentions can cause problems in the future and lessons to be learned for future projects.
---
Remote Connection:
https://ubc.zoom.us/j/63289130300?pwd=d7gDjPoLofSXQZW65395ZlOul0Z4IM.1
Meeting ID: 632 8913 0300
Passcode: 905770<br/><br/>Refreshments available 15min before the colloquium. BYOM- Bring your own mug!<br/><br/>______________________________<br/><br/>Detailed information available can be found at <a href='https://www.triumf.ca/research-program/lectures-conferences/upcoming-seminars-lectures'>https://www.triumf.ca/research-program/lectures-conferences/upcoming-seminars-lectures</a> <br/><br/>Date/Time: Wed Apr 02 2025 at 14:30<br/><br/>Location: ISAC II Conf. Room<br/><br/>Speaker: Prof. Vincenzo Patera (Sapienza Universita di Roma)<br/><br/>Title: Innovative particle beams for FLASH radiotherapy<br/><br/>Abstract: Over the past two decades, advances in cancer radiotherapy have driven continuous technological upgrades in the particle beams used. This trend has accelerated with the emergence of the FLASH effect, which requires ultra-high dose rate irradiation. FLASH radiotherapy has demonstrated a reduction in toxicity to normal tissue while maintaining tumor control in pre-clinical studies, compared to conventional dose rate radiotherapy.
Following the discovery of the FLASH effect, research into its biological mechanisms and potential clinical translation has spurred unprecedented innovation in radiotherapy beam technology. This seminar will present the key aspects of FLASH radiotherapy, along with an overview of the proton, electron, photon, and ion FLASH facilities currently in operation. Additionally, emerging future options, such as Very High Energy Electrons, will be discussed.
Join Zoom Meeting:
https://ubc.zoom.us/j/68240079189?pwd=5Xx6bmg5b5S4dXg1pVxB8F5yXBai4i.1
Meeting ID: 682 4007 9189
Passcode: 391108
<br/><br/> <br/><br/>______________________________<br/><br/>Detailed information available can be found at <a href='https://www.triumf.ca/research-program/lectures-conferences/upcoming-seminars-lectures'>https://www.triumf.ca/research-program/lectures-conferences/upcoming-seminars-lectures</a> <br/><br/>Date/Time: Wed Apr 02 2025 at 13:00<br/><br/>Location: Theory Room<br/><br/>Speaker: Tong Ou (U Chicago)<br/><br/>Title: Towards Real-Time Lattice Simulation of Baryogenesis<br/><br/>Abstract: Baryogenesis is a dynamical out-of-equilibrium process generating the baryon asymmetry of the Universe. Focusing on the mechanism of electroweak baryogenesis, where baryon number is generated through CP-violating scattering of the fermions with the bubble wall during a first-order electroweak phase transition, perturbative calculations for the relevant processes are known to suffer from various issues, motivating the development of lattice calculations. However, conventional lattice calculations based on Euclidean formulation in the imaginary time domain are limited to equilibrium dynamics. Real-time simulation is necessary for proper understanding of the out-of-equilibrium dynamics of baryogenesis. In our work, we focus on the fermion scattering process during electroweak baryogenesis, and develop a real-time simulation framework by mapping the fermions onto a spin system. Using tensor network methods, we perform numerical simulations and systematically analyze lattice artifacts. Our studies take a first step towards real-time simulation of baryogenesis, and are expected to shed light on real-time non-perturbative calculations of general out-of-equilibrium processes in the early Universe.
---
Remote Connection:
https://ubc.zoom.us/j/62535266430?pwd=THVrUzhMdjAyVkNnM05OUWRZYUppZz09
Meeting ID: 625 3526 6430
Passcode: 307647<br/><br/>.<br/><br/>______________________________<br/><br/>Detailed information available can be found at <a href='https://www.triumf.ca/research-program/lectures-conferences/upcoming-seminars-lectures'>https://www.triumf.ca/research-program/lectures-conferences/upcoming-seminars-lectures</a> <br/><br/>Date/Time: Mon Mar 31 2025 at 13:00<br/><br/>Location: Theory Room/Hybrid<br/><br/>Speaker: Anindita Maiti (Perimeter Institute)<br/><br/>Title: A Wilsonian RG framework for Regression Tasks in Supervised Learning (Early Career Talks)<br/><br/>Abstract:
Abstract: The performance of machine learning (ML) models fundamentally hinges on their ability to discriminate between relevant and irrelevant features in data. We introduce a first-of-its-kind Wilsonian RG framework to analyze the predictions of overparameterized neural networks (NN), which are models characterized by an excess of parameters relative to the complexity of the task. These networks, trained via supervised learning, are known to produce noisy outputs in regression tasks. In our formulation, irrelevant features within the data are systematically coarse-grained through momentum shell RG, inducing an RG flow that governs the evolution of noise in the predictions. When the irrelevant features follow a Gaussian distribution, this RG flow exhibits universality across different NN architectures. In contrast, non-Gaussian features give rise to more intricate, data-dependent RG flows. This approach reveals novel behaviors in NNs that have eluded conventional ML methods. By advancing beyond philosophical analogies between RG and ML, our framework offers a field theory-based methodology for understanding feature learning. This talk is based on the paper https://arxiv.org/abs/2405.06008.
---
Remote access: https://ubc.zoom.us/j/64575929910?pwd=QAb0SbY9Zi5iqxSB4Dfa1N9wGBXBqQ.1
Meeting ID: 645 7592 9910
Passcode: 892489
<br/><br/>A Wilsonian RG framework for Regression Tasks in Supervised Learning (Early Career Talks)
Early Career Talks (ECT)
Coffee and cookies available 15min before. BYO mug/cup
Remote access: https://ubc.zoom.us/j/61921073667?pwd=4Drp97meGJ3yq4Ro6k6LaoDncvJXaS.1
Meeting ID: 619 2107 3667
Passcode: 609573
Coffee and cookies available 15min before. BYO mug/cup<br/><br/>______________________________<br/><br/>Detailed information available can be found at <a href='https://www.triumf.ca/research-program/lectures-conferences/upcoming-seminars-lectures'>https://www.triumf.ca/research-program/lectures-conferences/upcoming-seminars-lectures</a> <br/><br/>