Search
Results
-
Towards developing a robust and generalizable brain computer interface
(2023)Recent advances in brain computer interfaces (BCI) have been shown to enable control of external prosthesis, restore paralyzed limb movements and to improve communication. In the field of communication, electrocorticography (ECoG), has emerged as one of the most promising techniques, especially when used together with deep learning. However, among the main challenges of this method are the low signal to noise ratio and the difficulty to ...Master Thesis -
Contextual Bandit Optimization with Pre-Trained Neural Networks
(2023)Bandit optimization is a difficult problem, especially if the reward model is high-dimensional. When rewards are modeled by neural networks, sublinear regret has only been shown under strong assumptions, usually when the network is extremely wide. In this thesis, we investigate how pre-training can help us in the regime of smaller models. We consider a stochastic contextual bandit with the rewards modeled by a multi-layer neural network. ...Master Thesis -
Assessment of Rock Mass Structures and Stability of the Rock-cut Tomb TT217 at Deir el-Medina (Luxor, Egypt)
(2023)The ancient Theban Tomb TT217, located in the Western Necropolis of the workers’ village Deir el-Medina, Luxor, Egypt, is carved into the densely fractured rock mass of the Thebes Limestone Formation and Esna Shale Formation. Already in antiquity, structurally controlled failure endangered the preservation of this UNESCO World Heritage Site and future rockfall cannot be ruled out. To understand the influence of the rock mass structures ...Master Thesis -
Subexponential Attacks on Variational LPN
(2023)Our Thesis presents a novel algebraic approach to solving the deci- sional Learning Parity with Noise (LPN) problem and hence the search variant as well. We studied different variations of the LPN problem regarding different distributions of the noise vector, as well as different fields over which the problem can be stated. We are offering two algorithms. One of them reduces the necessary sample size to 2n at the cost of slightly increasing ...Master Thesis -
Closed-Loop Identification of Sparse Networked Systems
(2023)In this project, we extend the application of two typical closed-loop identification methods, namely the dual-Youla parameterization and the two-stage identification, from identifying LTI SISO systems to MIMO systems while considering the predefined sparsity structure in the system. Both extended methods result in a quadratic program that consists of 1) a quadratic cost function that minimizes the model fitting error, and 2) a group of ...Master Thesis -
NOCS-Optimal Distributed Transactions and Beyond
(2023)This thesis aims to analyze and improve the current landscape of dis- tributed transactions. Specifically, we examine the state-of-the-art design principle, NOCS, and its limitations. Firstly, we test its upper bound, by implementing and testing a new algorithm, Eiger-PORT+ (EP+), which pushes it further to Transactional Causal Consistency with Convergence (TCCv). Our experiments show that EP+ outperforms the state of the art in throughput, ...Master Thesis -
Full-System Evaluation of the sPIN In-Network-Compute Architecture
(2023)In-network-computing with SmartNICs is gaining popularity in high-performance networking for their ability to offload packet processing tasks from the CPU and their latency advantage thanks to the proximity to the network traffic without having to go through PCIe. The sPIN in-network-computing paradigm developed at ETH Zürich aims to provide a programming model for developers to build high-performance packet processing routines for on-path ...Master Thesis -
Exploring heterogeneous OS architecture
(2023)In recent years the computer systems have been getting increasingly complex. In particular, these systems have increasingly heterogeneous cores. These heterogeneous cores often have a different view of shared memory, where they might see the same memory at different addresses. We show that software can run across cores with such a view of memory. We show a design of how we transform the addresses in a capability based system. This design ...Master Thesis -
Debuglets: an Inter-Domain Telemetry and Debugging Framework
(2023)After decades since the coming of the Internet, debugging issues between different domains remains a challenge. Only rudimental primitives are usually available to other network participants, such as the ICMP-based ping command, or the already-running applications. Even if some sophisticated tools have been built over the years, they are not always sufficiently powerful to quickly experiment, exclude, or confirm hypotheses about an issue. ...Master Thesis -
Hypervisor for a Multi-Tenant FPGA Heterogeneous System
(2023)Field-Programmable Gate Arrayss (FPGAs) are deployed widely by cloud providers, either as accelerators in their infrastructure or as infrastructure as a service. Coyote is an FPGA “OS” that simplifies the programming of the FPGA logic by providing similar abstractions as an traditional OS, such as virtual memory, physical memory management, a portable execution environment, and more. Coyote uses its Direct Memory Access (DMA) engines to ...Master Thesis