Tingwei Meng (Amazon Robotics): In-Context Operator Learning and Its Application in Financial Math
Please note this event occurred in the past.
April 24, 2026 11:00 am - 12:00 pm ET
LGRT 1685
Abstract
Abstract:
In this talk, I will present the In-Context Operator Network (ICON), a
project from my postdoctoral research. ICON leverages a single neural
network to solve diverse solution operators for applied math problems,
including ODEs, PDEs, and mean-field control problems, through
in-context learning. I will discuss the neural network architecture,
training methodology, and a follow-up application to optimal control
in financial mathematics. This is a collaboration with Liu Yang,
Siting Liu, Stanley Osher, Moritz Voss, Nils Detering, Giulio Farolfi,
and Georg Menz.
project from my postdoctoral research. ICON leverages a single neural
network to solve diverse solution operators for applied math problems,
including ODEs, PDEs, and mean-field control problems, through
in-context learning. I will discuss the neural network architecture,
training methodology, and a follow-up application to optimal control
in financial mathematics. This is a collaboration with Liu Yang,
Siting Liu, Stanley Osher, Moritz Voss, Nils Detering, Giulio Farolfi,
and Georg Menz.
Bio:
Tingwei Meng is an Applied Scientist at Amazon Robotics. She obtained
her Ph.D. in Applied Mathematics from Brown University and was
previously a Hedrick Assistant Adjunct Professor at UCLA. Her work
focuses on machine learning and optimization for high-dimensional
systems, with applications in scientific computing and robotics.
her Ph.D. in Applied Mathematics from Brown University and was
previously a Hedrick Assistant Adjunct Professor at UCLA. Her work
focuses on machine learning and optimization for high-dimensional
systems, with applications in scientific computing and robotics.