Wojciech Samek

Accurate and Robust Neural Networks for Security Related Applications Exampled by Face Morphing Attacks

2018 · arXiv (Cornell University)

Artificial neural networks tend to learn only what they need for a task. A manipulation of the training data can counter this phenomenon. In this paper, we study the effect of different alterations of the …

Black-Box Decision based Adversarial Attack with Symmetric α-stable Distribution

2019

Developing techniques for adversarial attack and defense is an important research field for establishing reliable machine learning and its applications. Many existing methods employ Gaussian random variables for exploring the data space to find the …

Shortcomings of Top-Down Randomization-Based Sanity Checks for Evaluations of Deep Neural Network Explanations

2022 · arXiv (Cornell University)

While the evaluation of explanations is an important step towards trustworthy models, it needs to be done carefully, and the employed metrics need to be well-understood. Specifically model randomization testing is often overestimated and regarded …

From attribution maps to human-understandable explanations through Concept Relevance Propagation

2023 · Nature Machine Intelligence

Abstract The field of explainable artificial intelligence (XAI) aims to bring transparency to today’s powerful but opaque deep learning models. While local XAI methods explain individual predictions in the form of attribution maps, thereby identifying …

PURE: Turning Polysemantic Neurons Into Pure Features by Identifying Relevant Circuits

2024 · arXiv (Cornell University)

The field of mechanistic interpretability aims to study the role of individual neurons in Deep Neural Networks. Single neurons, however, have the capability to act polysemantically and encode for multiple (unrelated) features, which renders their …

Leveraging Sparsity for Privacy in Collaborative Inference

2026

Collaborative inference (CI) is hampered by high communication costs and privacy risks, with existing defenses often forcing a trade-off between efficiency and formal privacy guarantees. In this work, we present a framework that leverages activation …