Upcoming Events - Feb 9 - Feb 15, 2025

Stay informed about the latest events happening within the CEMSE Division. Here is what is coming up next week:

KAUST-CEMSE-ECE-Graduate-Seminar-Dhanu-Chettri-Development-of-a-Four-Quadrant-Gallium-Oxide-Bidirectional-Switch

Development of a Four-Quadrant Gallium Oxide Bidirectional Switch

-

B9, L2, R2325

With the relentless advancement of electronics and the growing adoption of electric vehicles and renewable energy systems, the need for bidirectional control of electricity becomes increasingly crucial. This demand is further amplified by the necessity for high-efficiency, high-power devices that are both smaller and lighter.
KAUST-CEMSE-AMCS-Bridging-Optimal-Control-and-Modern-AI-Techniques-Tingwei-Meng

Bridging Optimal Control and Modern AI Techniques

-

Auditorium between 2 and 3

The convergence of artificial intelligence (AI) and applied mathematics offers transformative potential for addressing challenges in scientific computing and control. In this talk, I will highlight my research at the intersection of optimal control and modern AI methods, focusing on the development of scalable, robust, and efficient algorithms to tackle high-dimensional systems with uncertainty.
KAUST-CEMSE-CS-Seminar-Aleksei-Kalinov

Surface Tracking for Large-scale Dynamic Foams and Fluids

-

B1, L2, R2202

Foams produce the most lively phenomena that we encounter in our every-day lives. Think about starch bubbles rising in a rice cooker or soap bubbles filling a bathtub. Simulating such scenarios at a large scale is a challenge, due to constant topology changes, high frequency of degenerate geometry configurations, and huge storage and efficiency requirements. In my talk I will present a new multi-material surface tracking tool, designed to robustly handle large-scale deforming geometry, while preserving surface features. I will show how the algorithm combines explicit and implicit geometry representations, enjoying the advantages of both approaches. Finally, we will discuss how this tool enables extremely detailed simulations and exploration of large scale foams.

Spatial Dataflow is all you need!

-

B9 L2 R2325

Jian Weng's research, inspired by spatial dataflow, proposes a unified hardware design and implementation framework that leverages a familiar software programming paradigm, enabling concise and efficient designs while significantly reducing code complexity and achieving near-handcrafted performance, with future applications in ML inference systems targeting memory bandwidth limitations.

Mapping Properties of Neural Networks and Inverse Problems

-

B1 L4 R4214

This talk will explore the injectivity and universal approximation properties of neural networks, focusing on one-to-one mappings and the approximation of probability measures using compositions of invertible flow networks and injective layers, with applications in inverse problems, imaging, and generative models.
KAUST-CEMSE-AMCS-STAT-Graduate-Seminar-Aelson-Sobral-infinity-laplacian

Recent Advances in Infinity Laplacian Equations

-

B9, L2, R2325

The infinity Laplacian is a highly degenerate, nonlinear, second-order operator that arises in the study of optimal Lipschitz extensions and tug-of-war games, where two players alternately choose directions to move. In this presentation, we will provide a comprehensive motivation for key aspects of the infinity Laplacian and report on recent advances in the qualitative properties of its evolutionary counterpart.