Wieland BrendelinBethgelabAccurate, reliable and fast robustness evaluationA novel class of adversarial attacks that can determine the robustness of machine learning models with greater accuracy and reliability.6 min read·Jan 16, 2020----
Wieland BrendelinBethgelabNeural Networks seem to follow a puzzlingly simple strategy to classify imagesThis explains a range of observations, leads to more interpretable models and suggests novel avenues towards human-like perception.7 min read·Feb 6, 2019--21--21
Wieland BrendelinBethgelabResults of the NIPS Adversarial Vision Challenge 2018The winners of this year’s NIPS Adversarial Vision Challenge have been determined. This year the competition focused on a real-world…5 min read·Nov 9, 2018----
Wieland BrendelinLayer7 AIDetecting overhyped Artificial Intelligence solutionsAs a non-expert, it’s often difficult to detect whether a commercial AI solution is legitimate or bogus. Here are a few tips to find out.5 min read·Nov 7, 2018----
Wieland BrendelinBethgelabAI still fails on robust handwritten digit recognition (and how to fix it)Deep Learning has been praised to solve everything from self-driving cars to the world-climate. And yet, deep neural networks (the…5 min read·Sep 21, 2018--1--1
Wieland BrendelReading list for the NIPS 2018 Adversarial Vision ChallengeThe Adversarial Vision Challenge at NIPS 2018 has recently started and will run until November 1st. A key ingredient of the challenge is…3 min read·Jul 25, 2018----