32 citations · 106 across the 33 of their papers we have counts for
39 papers
Tackle CSM in JPEG Steganalysis with Data Adaptation
Rony Abecidan, Vincent Itier, Jérémie Boulanger +2
Steganalysis models excel on benchmark datasets but struggle in the wild when analyzed images are produced by a processing pipeline unseen during training. This problem known as Co…
Intersectional Fairness via Mixed-Integer Optimization
Jiří Němeček, Mark Kozdoba, Illia Kryvoviaz +2
The deployment of Artificial Intelligence in high-risk domains, such as finance and healthcare, necessitates models that are both fair and transparent. While regulatory frameworks,…
Targeted Pooled Latent-Space Steganalysis Applied to Generative Steganography, with a Fix
Etienne Levecque, Aurélien Noirault, Tomáš Pevn{ý} +3
Steganographic schemes dedicated to generated images modify the seed vector in the latent space to embed a message. Whereas most steganalysis methods attempt to detect the embeddin…
Distillation of a tractable model from the VQ-VAE
Armin Hadžić, Milan Papez, Tomáš Pevný
Deep generative models with discrete latent space, such as the Vector-Quantized Variational Autoencoder (VQ-VAE), offer excellent data generation capabilities, but, due to the larg…
Sparse Probabilistic Graph Circuits
Martin Rektoris, Milan Papež, Václav Šmídl +1
Deep generative models (DGMs) for graphs achieve impressively high expressive power thanks to very efficient and scalable neural networks. However, these networks contain non-linea…
Probabilistic Graph Circuits: Deep Generative Models for Tractable Probabilistic Inference over Graphs
Milan Papež, Martin Rektoris, Václav Šmídl +1
Deep generative models (DGMs) have recently demonstrated remarkable success in capturing complex probability distributions over graphs. Although their excellent performance is attr…
GraphSPNs: Sum-Product Networks Benefit From Canonical Orderings
Milan Papež, Martin Rektoris, Václav Šmídl +1
Deep generative models have recently made a remarkable progress in capturing complex probability distributions over graphs. However, they are intractable and thus unable to answer…
Sum-Product-Set Networks: Deep Tractable Models for Tree-Structured Graphs
Milan Papež, Martin Rektoris, Tomáš Pevný +1
Daily internet communication relies heavily on tree-structured graphs, embodied by popular data formats such as XML and JSON. However, many recent generative (probabilistic) models…
Malicious Internet Entity Detection Using Local Graph Inference
Simon Mandlik, Tomas Pevny, Vaclav Smidl +1
Detection of malicious behavior in a large network is a challenging problem for machine learning in computer security, since it requires a model with high expressive power and scal…
Generating Likely Counterfactuals Using Sum-Product Networks
Jiri Nemecek, Tomas Pevny, Jakub Marecek
The need to explain decisions made by AI systems is driven by both recent regulation and user demand. The decisions are often explainable only post hoc. In counterfactual explanati…
Optimize Planning Heuristics to Rank, not to Estimate Cost-to-Goal
Leah Chrestien, Tomás Pevný, Stefan Edelkamp +1
In imitation learning for planning, parameters of heuristic functions are optimized against a set of solved problem instances. This work revisits the necessary and sufficient condi…
Leveraging Data Geometry to Mitigate CSM in Steganalysis
Rony Abecidan, Vincent Itier, Jérémie Boulanger +2
In operational scenarios, steganographers use sets of covers from various sensors and processing pipelines that differ significantly from those used by researchers to train stegana…
Improving the Validity of Decision Trees as Explanations
Jiri Nemecek, Tomas Pevny, Jakub Marecek
In classification and forecasting with tabular data, one often utilizes tree-based models. Those can be competitive with deep neural networks on tabular data and, under some condit…
Is AUC the best measure for practical comparison of anomaly detectors?
Vít Škvára, Tomáš Pevný, Václav Šmídl
The area under receiver operating characteristics (AUC) is the standard measure for comparison of anomaly detectors. Its advantage is in providing a scalar number that allows a nat…
Using Set Covering to Generate Databases for Holistic Steganalysis
Rony Abecidan, Vincent Itier, Jérémie Boulanger +2
Within an operational framework, covers used by a steganographer are likely to come from different sensors and different processing pipelines than the ones used by researchers for…
A Differentiable Loss Function for Learning Heuristics in A*
Leah Chrestien, Tomas Pevny, Antonin Komenda +1
Optimization of heuristic functions for the A* algorithm, realized by deep neural networks, is usually done by minimizing square root loss of estimate of the cost to goal values. T…
Explaining Classifiers Trained on Raw Hierarchical Multiple-Instance Data
Tomáš Pevný, Viliam Lisý, Branislav Bošanský +2
Learning from raw data input, thus limiting the need for feature engineering, is a component of many successful applications of machine learning methods in various domains. While m…
Heuristic Search Planning with Deep Neural Networks using Imitation, Attention and Curriculum Learning
Leah Chrestien, Tomas Pevny, Antonin Komenda +1
Learning a well-informed heuristic function for hard task planning domains is an elusive problem. Although there are known neural network architectures to represent such heuristic…
Fitting large mixture models using stochastic component selection
Milan Papež, Tomáš Pevný, Václav Šmídl
Traditional methods for unsupervised learning of finite mixture models require to evaluate the likelihood of all components of the mixture. This becomes computationally prohibitive…
When Should You Defend Your Classifier -- A Game-theoretical Analysis of Countermeasures against Adversarial Examples
Maximilian Samsinger, Florian Merkle, Pascal Schöttle +1
Adversarial machine learning, i.e., increasing the robustness of machine learning algorithms against so-called adversarial examples, is now an established field. Yet, newly propose…
Mill.jl and JsonGrinder.jl: automated differentiable feature extraction for learning from raw JSON data
Simon Mandlik, Matej Racinsky, Viliam Lisy +1
Learning from raw data input, thus limiting the need for manual feature engineering, is one of the key components of many successful applications of machine learning methods. While…
Mapping the Internet: Modelling Entity Interactions in Complex Heterogeneous Networks
Simon Mandlik, Tomas Pevny
Even though machine learning algorithms already play a significant role in data science, many current methods pose unrealistic assumptions on input data. The application of such me…
Comparison of Anomaly Detectors: Context Matters
Vít Škvára, Jan Franců, Matěj Zorek +2
Deep generative models are challenging the classical methods in the field of anomaly detection nowadays. Every new method provides evidence of outperforming its predecessors, often…
Neural Power Units
Niklas Heim, Tomáš Pevný, Václav Šmídl
Conventional Neural Networks can approximate simple arithmetic operations, but fail to generalize beyond the range of numbers that were seen during training. Neural Arithmetic Unit…