Do you ever wonder what teachers and PhD'ers do aside from education? Then visit the colloquium! During this activity employees of all different research groups (Data Science, Digital Security and Software Science) will tell something about their research. In three stages you'll learn more about the research of Tom Claessen, Joost Rijneveld and Frits Vaandrager. At the end we'll discuss everything while enjoying a drink.
Below you can find more information on the talks.
Tom Claassen: A Brief Introduction to Causal Discovery.
Discovering causal relations from real-world data is one of the key goals of modern scientific research. Despite recent advances, many researchers are still not familiar with, and often even skeptical about, the possibilities of finding such relations from purely observational data. In this talk we will have a look at the famous ‘Simpson’s Paradox' to see why it is so important to at least try to make a principled distinction between correlation and causation, and what you would need to do that. We will discuss the key idea behind algorithms that try to find such cause-effect relations, so you have a chance to form your own opinion on whether you agree this actually makes sense or not. We will close with some example applications of this increasingly important area of machine learning, and currently ‘hot' topics in causal research.
Joost Rijneveld: Post-quantum cryptography
Cryptography plays a crucial role in our everyday interactions with the digital systems around us. For this, we rely on so-called 'hard problems': as long as these mathematical problems cannot be solved efficiently, our messages cannot be decrypted and our signatures cannot be forged. But what if an adversary gains access to a large-scale quantum computer, and all of our problems actually turn out to be easy? And what if we're still stuck with the 'classical' servers we bought in 2015?
In this talk I will give a brief glimpse into a new era of cryptography, where all integers have been factored (for large values of 'all') and RSA has finally been carried to its grave. As an illustration, I describe a promising new class of signature schemes: hash-based signatures.
Frits Vaandrager: Active Automata Learning
In this presentation, I will review the basic theory of active learning of state machines and recent applications in which this theory was used to learn models of (and find bugs in) smart cards, implementations of network protocols, and embedded systems controllers. I discuss some recent results and outline research challenges.