Edward Liem: On Extending Abductive Explanations
In my last colloquium, I discussed the importance of explainable AI (XAI) and steps towards explaining ML classifiers. In particular, I showed how to obtain abductive explanations (AXps), how certain AXps may be illogical from the user’s perspective, and how to omit AXps based on given criterion. Moreover, I illustrated how one could extract an …continue reading
Menno Bartels: Two Challenges for Symmetry Reduction
Within the theory of model checking one of the techniques to combat the state space explosion is symmetry reduction. The main idea is to combine symmetric states, resulting in a reduced state space. This technique poses two main challenges: symmetry detection and symmetry reduction. In the previous talks we presented a technique that relies on …continue reading
Tar van Krieken: Symbolic reachability using Quasi Decision Diagrams
We previously introduced Quasi Decision Diagrams (QDDs) which act as a replacement of Binary Decision Diagrams. In this talk, we discuss how QDDs can be used for reachability analysis, and we highlight the challenges that arise. We also propose solutions for these challenges, which remain to be evaluated within the overall reachability pipeline.
Myrthe Spronck: Template formulae and strong fairness
I have previously presented template modal μ-calculus formulae for capturing a variety of liveness properties under a number of different completion criteria. We have since improved on this work: we have designed a formula for strong fairness that is polynomial in the number of non-blockable actions, rather than exponential. Additionally, Lise Arendsen has formalized the …continue reading
Lars van den Haak: Scalable Deductive Verification of Data-Level Parallel Programs
We make reasoning about potentially overlapping arrays easier, by providing specification constructs to indicate and verify that two arrays are not aliases, or that they are immutable, so they can be modelled as mathematical sequences. All our techniques are implemented in the VerCors program verifier. We illustrate how our techniques improve scalability through a large …continue reading
Thomas Neele: A logical view on paradises in parity games, with some SMT help
A parity game is an infinite-duration game between two players who pass a token around in a directed graph. The problem of “solving” a parity game (deciding which player wins a given node) is an interesting problem since it is in UP and co-UP, but not known to be in P. In a binary parity …continue reading
Valentina Castiglioni: Return to the origins: equivalences, approximations and metrics for nondeterministic probabilistic processes
In this talk, I will give an overview of the main instruments that I encountered during my PhD to reason formally on the behaviour of processes exhibiting both nondeterministic and probabilistic behaviour. We will talk about behavioural equivalences, their approximated versions and behavioural metrics. If time permits, we will also discuss the logical characterisations of …continue reading
Sebastián Betancourt: Monitoring under uncertainty with DisTL
We have been investigating monitoring with STARK for more than a year now. Following our first publication on the topic, I would like to take a step back and present our work from a high-level perspective. I will provide some intuition behind DisTL and its quantitative semantics, introduce the monitors we have developed and their …continue reading
Tom Verhoeff: The Picture-Hanging Puzzles: Recent Progress
The classic picture-hanging puzzle asks for a way to wrap a wire around two nails such that the picture hangs as long as no nails are removed, but falls as soon as any one nail is removed. Mathematicians have generalized this puzzle in several steps: • The picture hangs on N nails and falls when …continue reading
Arend Rensink: Visualising CTL Witnesses and Counterexamples
One of the advantages of LTL over CTL is that the notion of a counterexample is easy to grasp, visualise and process: it is a trace that violates the property at hand. In this paper we propose a notion of evidence for CTL properties on explicit-state models — which equally serves as witness for satisfied …continue reading
