no code implementations • 12 Sep 2023 • Tuan Pham Minh, Jayan Wijesingha, Daniel Kottke, Marek Herde, Denis Huseljic, Bernhard Sick, Michael Wachendorf, Thomas Esch
Since these initial labels are partially erroneous, we use active learning strategies to cost-efficiently refine the labels in the second step.
1 code implementation • 16 Jun 2023 • Lukas Rauch, Matthias Aßenmacher, Denis Huseljic, Moritz Wirth, Bernd Bischl, Bernhard Sick
Deep active learning (DAL) seeks to reduce annotation costs by enabling the model to actively query instance annotations from which it expects to learn the most.
1 code implementation • 5 Apr 2023 • Marek Herde, Denis Huseljic, Bernhard Sick
Training standard deep neural networks leads to subpar performances in such multi-annotator supervised learning settings.
1 code implementation • 12 Oct 2022 • Marek Herde, Zhixin Huang, Denis Huseljic, Daniel Kottke, Stephan Vogt, Bernhard Sick
Retraining deep neural networks when new data arrives is typically computationally expensive.
1 code implementation • 6 Oct 2022 • Denis Huseljic, Marek Herde, Mehmet Muejde, Bernhard Sick
In the field of deep learning based computer vision, the development of deep object detection has led to unique paradigms (e. g., two-stage or set-based) and architectures (e. g., Faster-RCNN or DETR) which enable outstanding performance on challenging benchmark datasets.
no code implementations • 23 Sep 2021 • Marek Herde, Denis Huseljic, Bernhard Sick, Adrian Calma
Therefore, we introduce a general real-world AL strategy as part of a learning cycle and use its elements, e. g., the query and annotator selection algorithm, to categorize about 60 real-world AL strategies.
1 code implementation • 4 May 2021 • Felix Möller, Diego Botache, Denis Huseljic, Florian Heidecker, Maarten Bieshaar, Bernhard Sick
For this purpose, we propose a novel approach that allows for the generation of out-of-distribution datasets based on a given in-distribution dataset.
1 code implementation • 2 Jun 2020 • Daniel Kottke, Marek Herde, Christoph Sandrock, Denis Huseljic, Georg Krempl, Bernhard Sick
Gathering labeled data to train well-performing machine learning models is one of the critical challenges in many applications.
no code implementations • 29 Jan 2019 • Daniel Kottke, Jim Schellinger, Denis Huseljic, Bernhard Sick
Hence, it is not possible to reliably estimate the performance of the classification system during learning and it is difficult to decide when the system fulfills the quality requirements (stopping criteria).