In computer vision, generative AI models are typically built for images, videos, and 3D objects. Recently, there has emerged a paradigm of neural fields, which unifies the representations of such types of data by parametrizing them via neural networks. In this thesis, we develop generative models for images, videos, and 3D scenes which treat the underlying data in such a form and explore the benefits which such a perspective provides.

The success of Generative Adversarial Networks (GANs) has resulted in unprecedented quality both for image generation and manipulation. Recent state-of-the-art GANs (e.g., the StyleGAN series) have demonstrated outstanding results in photo-realistic image generation. In this dissertation, we explore the latent space properties, including image manipulation, extraction of 3D properties, and performing various weakly supervised and unsupervised downstream tasks using StyleGAN and its derivative architectures.

In this dissertation, three different coupled systems of volume integral and hydrodynamic equations are formulated to analyze electromagnetic scattering from composite plasmonic structures consisting of dielectrics, metals, or semiconductors.

Leveraging on the unique solar light absorption behavior, chemical stability, and piezoelectric properties, nanostructured group-III-nitrides were utilized for implementing various renewable energy harvesting and conversion applications. By using different dimensional nanostructures, especially nanowires, and nanoporous membranes, this dissertation explored the practical applications in photocatalytic nitrogen reduction reaction, photocatalytic wastewater remediation, and flexible self-powered piezoelectric sensing.

We will present a comprehensive study of a finite volume method for inviscid and viscous flow fields at high and low speeds. Thereby, the results of a formal asymptotic low Mach number analysis are used to extend the validity of the numerical method from the simulation of compressible flow fields at transonic as well as supersonic speed to the low Mach number regime.

We present order conditions for various Patankar-type schemes as well as a new stability approach that examines the non-hyperbolic fixed points of the schemes for a general linear test problem. We formulate sufficient conditions for the stability of such non-hyperbolic fixed points and the local convergence of the numerical approximation towards the correct steady-state solution of the underlying test problem. To illustrate the theoretical results, we consider several members of the modified Patankar-type family within numerical experiments.

As one of the founding departments in the Hong Kong Polytechnic University (PolyU) eighty-five years ago, the Department of Mechanical Engineering (ME) has been actively involved in both fundamental and applied research to advance the frontier of knowledge to serve Hong Kong, its surrounding region and beyond. The research foci of the department are in four major areas under three Research Centres and a Research Consortium under the University Niche Area scheme.

Entrepreneurs continue to be the driver for economic development and innovation. Some startups invent brand new markets while other manage to enter markets crowded by existing large companies. In this seminar we will explore making critical early decisions starting from chaos and creating an exciting new business. We will gain insights on the value of learning by doing, prototyping, discussing tradeoffs between analysis, experimentation and scale.  We will also review courses offered by KAUST Entrepreneurship Center.

In this talk, we present theoretical asymptotic results for the nonparametric estimation of the conditional density of a scalar response variable Y given the explanatory X taking values in a finite-dimensional space when the sample of observations is considered as a sequence of dependent random variables.

Recent years have been pivotal in the field of Industrial Control Systems (ICS) security, with a large number of high-profile attacks exposing the lack of a design-for-security initiative in ICS.

In this talk, I will first discuss how my research democratizes the accelerator designs and unifies the hardware/software innovations by automating the accelerator design process under a unified programming paradigm. By taking advantage of the compiler’s awareness of the program behaviors that profit from hardware specialization, accelerators can be automatically synthesized by searching through a well-defined design space. These automatically designed accelerators achieve comparable cost/performance efficiency compared with prior handcrafted designs. In the rest of the talk, I will also cover how this work inspires me to develop techniques to accelerate emerging application domains by orders-of-magnitude speedup, including digital signal processing and DNN inference, and how I take advantage of this work to revolutionize the FPGA programming paradigm.

Vulnerability discovery can be challenging: many software packages, both open and closed-source projects, build on existing code from public software repositories to network drives, derived from earlier versions or related software packages. Even implemented protocols rely on such repositories. It is important to detect such copies of code when the original code contains a software vulnerability, especially one that is exploitable, as seen with flaws such as the bash vulnerability Shellshock or the SSL vulnerability Heartbleed.

In this talk we first provide an introduction to point processes, which are stochastic models for the occurrence of events in space and time. We then discuss the application of point processes to investigate the relationship between law enforcement drug seizures and accidental overdoses in Indianapolis. We will also discuss results from a field-experiment in Indianapolis where point process based harm indices were used to inform the distribution of addiction treatment information. 

In this presentation, we will cover both communication and sensing security from a broader perspective. Even though, more emphasis on PHY security is given, other security measures will also be covered for the sake of completeness and as a step towards cross-layer security and cognitive security vision.

Machine Learning (ML) tools have recently been adopted for a wide range of automated operations in optical networking, moving fundamental steps towards the paradigm of zero-touch infrastructures. One example of such tasks is estimating the Quality of transmission of a lightpath prior to its establishment, which is particularly challenging due to the non-linear characteristics of signal propagation in optical fibers and to the often-incomplete knowledge of equipment parameters. This talk provides an overview of the contribution of my research team in the field of ML-based lightpath QoT estimation, including transfer learning approaches for inter-domain model adaptation, active learning for model building with small-sized training dataset, quantification of prediction uncertainty, and adoption of Explainable AI framework to expose the internal decisional mechanisms of trained models.

Coffee Time: 15:30-16:00. It has been a long-standing debate that “Is deep learning alchemy or science? ”, since the success of deep learning mainly relies on various engineering design and tricks, and lack of theoretical foundation. Unfortunately, the underlying mechanism of deep learning is still mysterious, severely limiting its further development from both theoretical and application aspects.

Neuromorphic computing has emerged as a new and promising computing principle that emulates how human brains process information. The underlying spiking neural networks (SNNs) are well-known for having higher energy efficiency than artificial neural networks (ANNs). Neuromorphic systems enable highly parallel computation and reduce memory bandwidth limitations, making hardware performance scalable and sustainable given the ever-increasing complexities of artificial intelligence (AI). Inefficiency in the design of a neuromorphic system generally originates from redundant parameters, nonoptimized models, a lack of computing parallelism, and inefficient training algorithms. This dissertation aims to address these problems and propose effective solutions.

Abstract

This tutorial series will include:

Bioinformatics experts, Don’t miss the opportunity to collaborate with researchers and field professionals in the BioHackathonMENA2023 event. BioHackathon events involve a large number of people that meet on-site to discuss ideas and implement projects in a collaborative manner during intensive coding sessions.

In this work we focus our attention on distributed optimization problems in the context where the communication time between the server and the workers is non-negligible. We obtain novel methods supporting bidirectional compression (both from the server to the workers and vice versa) that enjoy new state-of-the-art theoretical communication complexity for convex and nonconvex problems.

The understanding and description of turbulent flows poses several different problems in which theory, modeling, and (numerical) experiments interact in a multidisciplinary way. In this talk we present some of the basic tools, focusing on the mathematical foundation of the models used to describe the larger scales of the incompressible motions at high Reynolds number. Recent perspectives and open problems are also discussed.

Wave functional materials are artificial materials that can control wave propagation as wish. In this talk, I will give a brief review on the progress of wave functional materials and reveal the secret behind the engineering of these materials to achieve desired properties.  In particular, I will focus on our contributions on metamaterials and metasurfaces.

In this talk we will explore how research in systems and distributed systems may improve the resilience to cyber attacks following 3 axes targeting mobile systems, distributed systems, and operating systems

We propose a structured latent ODE model that explicitly captures system input variations within its latent representation. Building on a static latent variable specification, our model learns (independent) stochastic factors of variation for each input to the system, thus separating the effects of the system inputs in the latent space. This approach provides actionable modeling through the controlled generation of time-series data for novel input combinations (or perturbations). Additionally, we propose a flexible approach for quantifying uncertainties, leveraging a quantile regression formulation.

The talk will present our efforts to develop the next-generation operational system for the Red Sea, as part of Aramco’s resolution toward the fourth industrial revolution. This integrated system has been built around state-of-the-art ocean-atmosphere-wave general circulation models that have been specifically developed for the Red Sea region and nested within the global weather systems.