|
Gabriela Petculescu
Louisiana Accelerator Center
University of Louisiana at Lafayette
Elastic Anisotropy in Additively Manufactured 316L Stainless Steel
The large parameter space in selective laser melting (SLM)
— an additive manufacturing (a.k.a., 3D printing) technology
— presents a challenge for property predictors. In this
seminar, a study of elastic and corrosion properties of SLM 316L
under constant 100 W laser power and variable laser speed (600 to
1200 mm/s) is presented. Property variations and their relation to
the change in material microstructure are analyzed. Resonant
ultrasound spectroscopy on mm-sized rectangular parallelepipeds was
used for measuring location-dependent elastic moduli. Complementary
pulse-echo measurements provided volume-integrated values for the
longitudinal elastic modulus. Cyclic potentiodynamic polarization
measurements were used to determine the corrosion potential, pitting
potential, repassivation current, and corrosion current density of
the fabricated samples. The microstructure was determined with
Electron Backscatter Diffraction. The range of properties observed,
understood through grain boundary density and misorientation
distribution analysis, reinforces the capabilities of materials
produced through SLM: the tuning of fabrication parameters enables
materials tailoring for optimal performance in specific applications. |
|
Sophie Middleton
Division of Physics, Mathematics and Astronomy
California Institute of Technology (Caltech)
Hidden Worlds and Forbidden Processes: Using High Intensity Lepton Beams to Explore New Physics
Despite its successes, the Standard (SM) is fundamentally
incomplete, failing to fully account for the matter-antimatter
asymmetry, the nature of dark matter, and the origin of neutrino
masses. To address these shortcomings, the “intensity
frontier” offers a powerful probe into new physics that
either manifests itself at energy scales far beyond the direct
reach of current colliders or that only feebly interacts with SM
particles. In this talk, I will explore how high-intensity
searches — specifically at B-factories and the upcoming
Mu2e experiment — are poised to revolutionize our
understanding of the dark sector and lepton flavor physics. The
talk will also highlight the growing synergy between AI and
fundamental discovery. A compelling solution to the dark matter
puzzle lies in a “dark sector” of particles that are
neutral under SM forces. Low-energy, high-intensity colliders
like
BaBar (at
SLAC) or Belle and Belle II (in Japan) provide a unique
laboratory for these searches. I will discuss recent
world-leading results from
BaBar
and how advanced machine learning architectures are essential for
identifying rare dark sector signatures. Plans to utilize Belle
II to leverage a rapidly growing dataset to explore mesogenesis, dark photons, axion-like particles (ALPs), and heavy neutral
lepton (HNL) scenarios at unprecedented sensitivities will also
be presented. Furthermore, I will detail the status of the Mu2e
experiment at Fermilab, which seeks to observe the coherent,
neutrino-less conversion of a muon into an electron. This
process represents a clear signal of charged lepton flavor
violation (CLFV), unobservably suppressed in the SM. As Mu2e
prepares for its first physics run in 2027, I will highlight the
on-going preparations for physics analysis including deployment
of AI-driven particle identification and background
characterization. These tools are enabling Mu2e to probe
effective mass scales up to 10,000 TeV.
|
|
Kayla DeHolton
Department of Physics
Penn State University
Neutrino Oscillations at the South Pole and Beyond
Despite being one of the most abundant known particles in the
universe, neutrinos remain an enigma within the Standard Model.
The past quarter-century has seen great experimental strides in
measuring the properties of neutrino oscillations. However, many
fundamental questions remain unanswered, some of which can be
probed through atmospheric neutrino oscillations. The IceCube
DeepCore detector at the South Pole has been collecting GeV-scale
neutrino data for the past decade, and currently provides
world-leading measurements of the neutrino oscillation parameters
using atmospheric neutrinos. IceCube's ability to measure these
parameters will improve even further with the IceCube Upgrade
currently being deployed in Antarctica. The IceCube Upgrade will
consist of 7 additional densely-instrumented strings with new
types of modules containing multiple PMTs, greatly increasing
detector performance for GeV-scale neutrinos. Furthermore, the
techniques developed for these measurements can be readily
applied to additional science priorities, such as atmospheric
neutrino oscillations with the DUNE far detectors, GeV
multi-messenger astrophysics with IceCube, and improvements to
open-source deep learning reconstruction frameworks used in the
neutrino telescope community.
|
|
Shawn Dubey
Department of Physics
Brown University
A Tour of Tools, Trends, and Tradeoffs in Machine Learning for High Energy Physics
Machine Learning (ML) has advanced rapidly over the past decade,
transforming the way data-intensive sciences are conducted. High
energy physics is distinctive among the physical sciences in its
early adoption and deployment of new ML methods, driven by
extreme data volumes, complex event structures, and stringent
real-time constraints. In this talk, I will trace the evolution
of these approaches through my own research trajectory, beginning
with the use of multilayer perceptrons and physics-motivated
feature engineering, and progressing toward modern techniques
that emphasize data curation, weak supervision, and
hardware-software co-design of ML algorithms. I will highlight
cutting edge methods and how they arise from the interaction
between models, data, and experimental constraints, and how this
perspective informs future research directions.
|
|
Frank Meier
Vossen Group
Duke University
Precision Flavor Physics in the Era of Artificial Intelligence
The Standard Model of particle physics (SM) is a powerful
theoretical framework. However, many fundamental questions like
the explanation for the large matter-antimatter asymmetry
observed in today's universe remain unanswered. Precision
measurements and indirect searches offer a promising path to
uncover new insights and potential signs of physics beyond the
SM. Recent advances in computational techniques now allow us to
exploit the large experimental datasets more effectively,
reaching unprecedented levels of precision. In my talk, I will
discuss how machine learning is reshaping these precision
measurements in flavor physics, with a focus on semileptonic
decays of heavy mesons. These decays play a central role in
determining fundamental parameters of the Standard Model and in
probing possible violations of lepton flavor universality. I will
describe how machine-learning-based reconstruction and
classification methods dramatically improve signal efficiency and
background suppression compared to traditional approaches. Using
examples from my work at the Belle II experiment, I will show how
these techniques have enabled more precise measurements.
Finally, I will discuss how these methods generalize beyond
flavor physics and outline future opportunities for applying
modern AI tools to a wide range of data-intensive problems across
experimental physics.here the particles are significantly
accelerated by the dissipation of the magnetic field associated
to a possible reconnection manifestation.
|
