About this Event

Deep Neural Networks (DNNs) draw their power from the representations they learn. In recent years, however, researchers have found that DNNs, while being incredibly effective in learning complex abstractions, also tend to be infected with artifacts, such as biases, Clever Hanses (CH), or Backdoors, due to spurious correlations inherent in the training data. So far, existing methods for uncovering such artifactual and malicious behavior in trained models focus on finding artifacts in the input data, which requires both availabilities of a data set and human intervention. In this paper, we introduce DORA (Data-agnOstic Representation Analysis): the first automatic data-agnostic method for the detection of potentially infected representations in Deep Neural Networks. We further show that contaminated representations found by DORA can be used to detect infected samples in any given dataset. We qualitatively and quantitatively evaluate the performance of our proposed method in both, controlled toy scenarios, and in real-world settings, where we demonstrate the benefit of DORA in safety-critical applications.

Speaker

Kirill Bykov is a doctoral student in Machine Learning at the Technische Universität Berlin and ATB, with a focus on Interpretable and Explainable AI. When asked about his work, He likes to answer that he investigates the vivid diversity of the internal abstractions and representations learned by machines to understand how they perceive the world.

Apart from academia, He is also am a passionate photographer (only if he had more time to do that), an ardent reader and I am aroused by writing.