1 code implementation • 24 May 2023 • Daniël Vos, Jelle Vos, Tianyu Li, Zekeriya Erkin, Sicco Verwer
By leveraging the better privacy-utility trade-off of PrivaTree we are able to train decision trees with significantly better robustness against backdoor attacks compared to regular decision trees and with meaningful theoretical guarantees.
1 code implementation • 30 Jan 2023 • Daniël Vos, Sicco Verwer
While there is generally a trade-off between the performance and interpretability of machine learning models, we find that OMDTs limited to a depth of 3 often perform close to the optimal limit.
1 code implementation • 22 Aug 2022 • Azqa Nadeem, Daniël Vos, Clinton Cao, Luca Pajola, Simon Dieck, Robert Baumgartner, Sicco Verwer
The security literature sometimes also fails to disentangle the role of the various stakeholders, e. g., by providing explanations to model users and designers while also exposing them to adversaries.
1 code implementation • 8 Jul 2022 • Clinton Cao, Annibale Panichella, Sicco Verwer, Agathe Blaise, Filippo Rebecchi
The first step for these machine learning pipelines is to pre-process the data before it is given to the machine learning algorithm.
1 code implementation • 4 Jul 2022 • Robert Baumgartner, Sicco Verwer
In this paper we propose a method to learn state machines from data streams using the count-min-sketch data structure to reduce memory requirements.
no code implementations • 28 Jun 2022 • Clinton Cao, Agathe Blaise, Sicco Verwer, Filippo Rebecchi
In this work, we propose an approach that learns state machine models to model the runtime behaviour of a cloud environment that runs multiple microservice applications.
1 code implementation • 24 Jun 2022 • Azqa Nadeem, Sicco Verwer
K-medoids or Partitioning Around Medoids (PAM) is commonly used to cluster sequences since it supports alignment-based distances, and the k-centers being actual data items helps with cluster interpretability.
no code implementations • 15 Jun 2022 • Dennis Mouwen, Sicco Verwer, Azqa Nadeem
We present a method to learn automaton models that are more robust to input modifications.
1 code implementation • 28 Mar 2022 • Sicco Verwer, Christian Hammerschmidt
We present the efficient implementations of probabilistic deterministic finite automaton learning methods available in FlexFringe.
1 code implementation • 25 Jan 2022 • Laurens Bliek, Paulo da Costa, Reza Refaei Afshar, Yingqian Zhang, Tom Catshoek, Daniël Vos, Sicco Verwer, Fynn Schmitt-Ulms, André Hottung, Tapan Shah, Meinolf Sellmann, Kevin Tierney, Carl Perreault-Lafleur, Caroline Leboeuf, Federico Bobbio, Justine Pepin, Warley Almeida Silva, Ricardo Gama, Hugo L. Fernandes, Martin Zaefferer, Manuel López-Ibáñez, Ekhine Irurozki
Overall, by organizing this competition we have introduced routing problems as an interesting problem setting for AI researchers.
no code implementations • 8 Sep 2021 • Daniël Vos, Sicco Verwer
Decision trees are a popular choice of explainable model, but just like neural networks, they suffer from adversarial examples.
1 code implementation • 6 Jul 2021 • Azqa Nadeem, Sicco Verwer, Shanchieh Jay Yang
We propose to automatically learn AGs based on actions observed through intrusion alerts, without prior expert knowledge.
1 code implementation • 8 Jun 2021 • Laurens Bliek, Arthur Guijt, Rickard Karlsson, Sicco Verwer, Mathijs de Weerdt
Surrogate algorithms such as Bayesian optimisation are especially designed for black-box optimisation problems with expensive objectives, such as hyperparameter tuning or simulation-based optimisation.
1 code implementation • 18 Dec 2020 • Daniël Vos, Sicco Verwer
We can use algorithms that take adversarial attacks into account to fit trees that are more robust.
no code implementations • 6 Nov 2020 • Rickard Karlsson, Laurens Bliek, Sicco Verwer, Mathijs de Weerdt
One method to solve expensive black-box optimization problems is to use a surrogate model that approximates the objective based on previous observed evaluations.
no code implementations • 8 Jun 2020 • Laurens Bliek, Arthur Guijt, Sicco Verwer, Mathijs de Weerdt
A challenging problem in both engineering and computer science is that of minimising a function for which we have no mathematical formulation available, that is expensive to evaluate, and that contains continuous and integer variables, for example in automatic algorithm configuration.
1 code implementation • 20 Nov 2019 • Laurens Bliek, Sicco Verwer, Mathijs de Weerdt
When a black-box optimization objective can only be evaluated with costly or noisy measurements, most standard optimization algorithms are unsuited to find the optimal solution.
no code implementations • 29 Oct 2019 • Qin Lin, Sicco Verwer, John Dolan
Imitation learning provides a way to automatically construct a controller by mimicking human behavior from data.
no code implementations • 28 Jul 2017 • Christian A. Hammerschmidt, Radu State, Sicco Verwer
We present an interactive version of an evidence-driven state-merging (EDSM) algorithm for learning variants of finite state automata.
no code implementations • 6 Jun 2017 • Alexis Linard, Rick Smetsers, Frits Vaandrager, Umar Waqas, Joost van Pinxten, Sicco Verwer
A classical problem in grammatical inference is to identify a deterministic finite automaton (DFA) from a set of positive and negative examples.
no code implementations • 24 May 2017 • Xiaoran Liu, Qin Lin, Sicco Verwer, Dmitri Jarnikov
This paper focuses on detecting anomalies in a digital video broadcasting (DVB) system from providers' perspective.
no code implementations • 21 Nov 2016 • Christian Albert Hammerschmidt, Sicco Verwer, Qin Lin, Radu State
Automaton models are often seen as interpretable models.
1 code implementation • 8 Nov 2016 • Rick Smetsers, Joshua Moerman, Mark Janssen, Sicco Verwer
An ongoing challenge for learning algorithms formulated in the Minimally Adequate Teacher framework is to efficiently obtain counterexamples.
Software Engineering
no code implementations • 6 Jan 2014 • Sicco Verwer, Yingqian Zhang, Qing Chuan Ye
Given the learned models, we propose two types of optimization methods: a black-box best-first search approach, and a novel white-box approach that maps learned models to integer linear programs (ILP) which can then be solved by any ILP-solver.
no code implementations • 26 Sep 2013 • Maurice Bruynooghe, Hendrik Blockeel, Bart Bogaerts, Broes De Cat, Stef De Pooter, Joachim Jansen, Anthony Labarre, Jan Ramon, Marc Denecker, Sicco Verwer
This paper provides a gentle introduction to problem solving with the IDP3 system.