CEMSE Weekly Updates - November 10, 2026

Stay informed about the upcoming events and the latest news from CEMSE.

Call for Event Registration

Register for Course "Navier-Stokes and Euler Equations, and Certain Geophysical Models" before November 14. This four-lecture course will take place on Nov. 16, 18, 23 and 25, and will be led by Prof. Edriss S. Titi. This course supports graduate students from CEMSE and PSE Divisions to provide mathematical description of geophysical fluid dynamics and address challenges facing the turbulence community, particularly in the development of turbulence closure models.

Upcoming Events

The Grid Impact of AI Data Centers: When Markets and Adversaries Shape Workloads

-

B9 R2325

This talk explores how markets and adversaries can shape AI data center workloads, and what that means for grid stability, using market and grid simulations, hardware experiments, and real-time validation.

Trustworthy AI Begins in the Data Center

-

B9 R2325

Debates on trustworthy AI focus on model bias, alignment, and explainability. Yet, trust in AI is only as strong as the security and resilience of the infrastructure it runs on.

Navier-Stokes and Euler Equations, and Certain Geophysical Models - Course Lecture 1

-

Location shared with registered participants

This course will cover two main topics: the Navier-Stokes and Euler equations, and the global regularity of certain geophysical flows.

Advancing Monocular Depth Estimation: High-Resolution and Amodal Understanding

-

B1 R2202

This thesis advances monocular depth estimation to develop efficient methods for recovering fine geometric details, such as object boundaries and thin structures, from high-resolution images while preserving scene-wide consistency and metric scale.

Adaptive Local Proper Orthogonal Decomposition via Graph Partitioning and Provable Error-Bounded Lossy Compression

-

B1 R3426

This talk introduces a spatially adaptive POD-based compression method combining hierarchical domain partitioning and local low-rank approximations for error-controlled, storage-efficient compression of spatially heterogeneous simulation data.

Robust and Efficient Methods for Nonlinear Electromagnetic Inversion: Regularization, Optimization, and Learning-Assisted Strategies

-

B3 R5209

This dissertation develops several physics-based and learning-assisted methods to improve the robustness, capability, and computational efficiency of nonlinear EM inversion.

Local Formulations of Nonlocal Problems: Dispersive Tsunamis, Nonlinear Schrödinger Systems, and Absorbing Boundaries

-

B2/3 A0215

This thesis develops local approximations to problems whose natural formulation is nonlocal in different manners: through elliptic structure, dispersive operators, and boundary conditions.

Navier-Stokes and Euler Equations, and Certain Geophysical Models - Course Lecture 2

-

Location shared with registered participants

This course will cover two main topics: the Navier-Stokes and Euler equations, and the global regularity of certain geophysical flows.

Toward Monolithic Power Integration: β-Ga2O3 Devices and Circuits

-

B3 R5209

This dissertation develops beta-phase gallium oxide devices and circuits for monolithic power integration.

Efficient Long-Form Video Understanding: Scaling Perception, Selecting Evidence, and Adapting Computation

-
This dissertation advances efficient long-form video understanding by developing methods that allocate limited computational resources across three stages: scaling video perception, selecting relevant visual evidence, and adapting reasoning effort to individual questions.

Nonlinear Optimal Control and Filtering beyond the HJB Equation

-

B9 R2325

A novel set of sufficient conditions - equivalent to the HJB equation for the solution of nonlinear optimal control problems.

Modeling the Visual World Across Space and Time: From Observation to Interaction

-
This dissertation develops methods to recover and model the visual world consistently across space and time from images and video, enabling both accurate understanding of scene geometry and coherent, event-driven video editing grounded in user intent.

Registration

Register for upcoming event