Pattern Recognition Resonators This project demonstrates Resonate-and-Fire (RF) neurons and unsupervised hebbian learning can be combined to build Spiking Neural Networks (SNNs) that are capable of recognizing (complex) patterns in time data. Results We have no proper evaluation comparing against other state-of-the-art networks yet, as this code was created during a neuromorphic hackathon organized by neurotum x Fortiss. The hyperparameters, input data and output data of the experimental networks are saved in save.

Motivation and Goals For the interdisciplinary project (IDP), Ruben and I worked on a portfolio optimization tool for the Chair of Financial Management and Capital Markets at TUM. Our motivation for creating an EtfOptimizer tool was as follows: ETFs are a low cost and comparatively low risk investment option with good returns Abundance of ETFs on the market: complicates investment decisions An optimization can: Take investor preferences into account Asset categories Investment amount Risk tolerance Preferred return We had the following goals in mind while designing and implementing the etfoptimizer tool:

Pwning Artemis for fun and profit During the course, “WebApplication Security”, we received the task to find security vulnerabilities within Artemis, TUM’s own learning platform, widely deployed in large-scale (programming) courses at the computer science faculty. Within this blogpost I want to shortly discuss the security vulnerabilities I found in a writeup style. If you are interested in more detail read the full report and the slides of the talk my colleague and friend Florian Freund and I held.

During my third semester at TUM, I switched my OS from Windows to Linux. This happened mostly out of curiosity and wanting to learn about Linux, but also because all the “cool geeks” seemed to use some version of Linux. I ended up with the Kubuntu distribution, which ships with the Plasma Desktop and KDE applications simply because it looked a bit like Windows. Ever since that (ca. end of 2018), I have been an enthusiastic user of Plasma.

Fine-Grained Authorization with Spring Boot Spring offers many methods for checking authorization. In this short blog post I will focus on checking authorization at method-level and at the level of individual users. Method-Level Authorization First, we need to enable @PreAuthorize and @PostAuthorize annotations which are required for checking method-level security, by adding @EnableGlobalMethodSecurity(prePostEnabled = true) to a @Configuration bean that extends GlobalMethodSecurityConfiguration [1]. @Configuration @EnableGlobalMethodSecurity(prePostEnabled = true) public class MethodSecurityConfig extends GlobalMethodSecurityConfiguration { } Role checking with @PreAuthorize Now we can make use of Spring Security’s DSL to check whether the roles are fulfilled at individual methods:

Securing a Spring Boot Web Application Recently I had the chance to work on the authentication and authorization mechanisms for a simple Delivery application that we had to build in the Advanced Topics of Software Engineering class at TUM. This post is intended merely as a write-up of the most important steps that were required to fulfill the task. As a consequence, not everything here will be according to best practices.